{"templates":[{"id":"00000000-0000-0000-0000-000000000012","name":"ComfyUI","slug":"comfyui","description":"Node-based AI image generation with FLUX.2 klein. Drag, drop, and connect nodes to build complex image pipelines on GPU.","docker_image":"ghcr.io/benpoliquin/templates/comfyui@sha256:36e27866a87033bf7313cdc00c763426c39b9e49f6f8dd6ab97998967eb774e8","requires_gpu":true,"min_gpu":"rtx-3090","min_vram_gb":16,"estimated_setup_seconds":300,"category":"image-generation","source_repo":"https://github.com/comfyanonymous/ComfyUI","billing_model":"per_minute","proxy_port":8188,"monthly_price_cents":null,"volume_gb":50,"base_image_type":"custom","is_active":true,"datasets":[],"gpu_options":[{"id":"b66b8e05-6068-4bb2-b166-e2902e4ca7db","template_id":"00000000-0000-0000-0000-000000000012","gpu_type":"rtx-3090","display_name":"RTX 3090","sell_price_per_hr_cents":25,"is_default":false,"sort_order":0,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":150,"available":true,"stock_level":"medium","providers":[{"provider":"vastai","sell_price_per_hr_cents":25,"available":1,"stock_level":"medium","region":"United States, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":29,"available":1,"stock_level":"medium","region":"Nevada, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":36,"available":1,"stock_level":"low","region":"Utah, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":66,"available":1,"stock_level":"low","region":"Ohio, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":28,"max_vcpus":96,"min_ram_gb":15,"max_ram_gb":126},{"id":"1fe614b3-225c-4bed-9cb2-7e4b1354a240","template_id":"00000000-0000-0000-0000-000000000012","gpu_type":"rtx-4090","display_name":"RTX 4090","sell_price_per_hr_cents":48,"is_default":true,"sort_order":1,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":150,"available":true,"stock_level":"medium","providers":[{"provider":"runpod","sell_price_per_hr_cents":48,"available":1,"stock_level":"low","region":"","cloud_tier":"COMMUNITY"},{"provider":"vastai","sell_price_per_hr_cents":57,"available":1,"stock_level":"medium","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":61,"available":1,"stock_level":"medium","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"Pennsylvania, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"British Columbia, CA","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":85,"available":1,"stock_level":"low","region":"Minnesota, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":12,"max_vcpus":256,"min_ram_gb":31,"max_ram_gb":126},{"id":"a34f0acd-d220-4a6b-819d-a66204336c46","template_id":"00000000-0000-0000-0000-000000000012","gpu_type":"l40s","display_name":"L40S","sell_price_per_hr_cents":153,"is_default":false,"sort_order":2,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":300,"available":true,"stock_level":"low","providers":[{"provider":"runpod","sell_price_per_hr_cents":153,"available":1,"stock_level":"low","region":"","cloud_tier":"SECURE"}],"vram_gb":48,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"2fae6de3-ed5c-4a87-be84-add8026050df","template_id":"00000000-0000-0000-0000-000000000012","gpu_type":"a100-80gb-sxm","display_name":"A100 80GB SXM","sell_price_per_hr_cents":176,"is_default":false,"sort_order":3,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":500,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":176,"available":1,"stock_level":"low","region":"Montana, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":187,"available":1,"stock_level":"low","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":195,"available":1,"stock_level":"low","region":"North Carolina, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":307,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":80,"min_vcpus":128,"max_vcpus":208,"min_ram_gb":126,"max_ram_gb":126}],"frontend_config":{"gpu_badges":{"rtx-3090":"Budget pick"},"gpu_guidance":{"l40s":"48 GB VRAM — Flux full precision, stacked ControlNets, complex multi-model workflows.","rtx-3090":"24 GB VRAM — runs SDXL and Flux fp8 comfortably. Great budget option.","rtx-4090":"24 GB VRAM — best cost-per-image. Fastest generation at this price point.","a100-80gb-sxm":"80 GB VRAM — AI video generation (Wan 2.1, HunyuanVideo), production batching."},"what_you_get":["ComfyUI node-based workflow editor","FLUX.2 klein starter model + ready-to-run official workflow","ComfyUI Manager for one-click extensions","JupyterLab for scripting and file access"],"env_var_fields":[{"key":"COMFYUI_MODEL","label":"Starter Model","options":[{"label":"FLUX.2 klein 4B — 16 GB download (recommended)","value":"flux2-klein"},{"label":"SDXL Base — 6.5 GB download, ~8 GB VRAM (most tutorials use this)","value":"sdxl"},{"label":"Flux.1 Schnell (fp8) — 17.2 GB download, ~12 GB VRAM (faster, 4 steps)","value":"flux-schnell"},{"label":"Both SDXL + Flux Schnell — 23.7 GB download (longest first boot)","value":"both"},{"label":"No model — start empty, download your own later","value":"none"}],"required":true,"placeholder":"flux2-klein","allow_custom":false}],"show_gpu_selector":true,"expanded_description":"A node-based AI image generation interface powered by FLUX.2 klein — the current open-weight quality leader (Apache 2.0, no token needed). The official klein workflow is bundled and opens ready-to-run. Or pick SDXL / Flux Schnell as your starter model instead. Build complex image pipelines by connecting nodes — text-to-image, image editing, ControlNet, LoRA, and more. Includes ComfyUI Manager for one-click extension installs and JupyterLab for scripting.","optional_env_var_fields":[{"key":"HF_TOKEN","label":"HuggingFace Token","required":false,"helper_text":"Required for gated models (e.g. FLUX). Get yours at huggingface.co/settings/tokens","placeholder":"hf_..."},{"key":"CIVITAI_TOKEN","label":"CivitAI Token","required":false,"helper_text":"Required for downloading models from CivitAI","placeholder":""}]},"addons":{"available":[{"id":"jupyter","name":"JupyterLab","description":"Browser-based Python IDE with GPU monitoring","port":8888,"access_label":"JupyterLab","access_method":"jupyter"},{"id":"code-server","name":"VS Code Server","description":"Full VS Code editor in your browser","port":8443,"access_label":"VS Code","access_method":"code_server"}],"defaults":["jupyter"]}},{"id":"00000000-0000-0000-0000-000000000014","name":"Faster Whisper","slug":"faster-whisper","description":"GPU-accelerated speech-to-text and text-to-speech. Transcribe audio up to 4x faster than OpenAI Whisper, plus natural English TTS with Kokoro, via an OpenAI-compatible API.","docker_image":"ghcr.io/benpoliquin/templates/faster-whisper@sha256:dfb1ee9ea605b952280aad3ec82d415537e98bf39070766e4bc950a3fcd2e14b","requires_gpu":true,"min_gpu":"rtx-a4000","min_vram_gb":4,"estimated_setup_seconds":60,"category":"audio","source_repo":"https://github.com/speaches-ai/speaches","billing_model":"per_minute","proxy_port":8000,"monthly_price_cents":null,"volume_gb":20,"base_image_type":"custom","is_active":true,"datasets":[],"gpu_options":[{"id":"225d0dd0-a164-41fc-8160-bd56b5aff8fd","template_id":"00000000-0000-0000-0000-000000000014","gpu_type":"rtx-a4000","display_name":"RTX A4000","sell_price_per_hr_cents":16,"is_default":true,"sort_order":0,"compute_type":"gpu","gpu_count":1,"default_volume_gb":20,"max_volume_gb":100,"available":true,"stock_level":"medium","providers":[{"provider":"vastai","sell_price_per_hr_cents":16,"available":1,"stock_level":"medium","region":"Utah, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":24,"available":1,"stock_level":"low","region":"","cloud_tier":"COMMUNITY"}],"vram_gb":16,"min_vcpus":32,"max_vcpus":32,"min_ram_gb":18,"max_ram_gb":18},{"id":"7ccdb359-1e24-4fdf-802f-3f782461ff98","template_id":"00000000-0000-0000-0000-000000000014","gpu_type":"rtx-3090","display_name":"RTX 3090","sell_price_per_hr_cents":25,"is_default":false,"sort_order":1,"compute_type":"gpu","gpu_count":1,"default_volume_gb":20,"max_volume_gb":100,"available":true,"stock_level":"medium","providers":[{"provider":"vastai","sell_price_per_hr_cents":25,"available":1,"stock_level":"medium","region":"United States, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":29,"available":1,"stock_level":"medium","region":"Nevada, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":36,"available":1,"stock_level":"low","region":"Utah, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":66,"available":1,"stock_level":"low","region":"Ohio, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":28,"max_vcpus":96,"min_ram_gb":15,"max_ram_gb":126},{"id":"ad89aa2f-5c80-49bc-9d26-b6e6a2178784","template_id":"00000000-0000-0000-0000-000000000014","gpu_type":"l4","display_name":"L4","sell_price_per_hr_cents":61,"is_default":false,"sort_order":2,"compute_type":"gpu","gpu_count":1,"default_volume_gb":20,"max_volume_gb":100,"available":true,"stock_level":"low","providers":[{"provider":"vastai","sell_price_per_hr_cents":61,"available":1,"stock_level":"low","region":"Utah, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":97,"available":1,"stock_level":"low","region":"","cloud_tier":"SECURE"}],"vram_gb":24,"min_vcpus":128,"max_vcpus":128,"min_ram_gb":25,"max_ram_gb":25},{"id":"edf2cbc8-74cc-4a30-b67b-40c44dc24999","template_id":"00000000-0000-0000-0000-000000000014","gpu_type":"rtx-4090","display_name":"RTX 4090","sell_price_per_hr_cents":48,"is_default":false,"sort_order":3,"compute_type":"gpu","gpu_count":1,"default_volume_gb":20,"max_volume_gb":100,"available":true,"stock_level":"medium","providers":[{"provider":"runpod","sell_price_per_hr_cents":48,"available":1,"stock_level":"low","region":"","cloud_tier":"COMMUNITY"},{"provider":"vastai","sell_price_per_hr_cents":57,"available":1,"stock_level":"medium","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":61,"available":1,"stock_level":"medium","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"Pennsylvania, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"British Columbia, CA","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":85,"available":1,"stock_level":"low","region":"Minnesota, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":12,"max_vcpus":256,"min_ram_gb":31,"max_ram_gb":126}],"frontend_config":{"gpu_guidance":{"l4":"24 GB VRAM — datacenter-grade reliability, good all-rounder.","rtx-3090":"24 GB VRAM — more compute power, faster transcription at a great price.","rtx-4090":"24 GB VRAM — fastest transcription, ideal for bulk processing.","rtx-a4000":"16 GB VRAM — cheapest option, handles all Whisper models easily."},"what_you_get":["Drag-and-drop transcription web UI","Natural English text-to-speech with Kokoro TTS","OpenAI-compatible API (/v1/audio/transcriptions, /v1/audio/speech)","99 language support with auto-detection","Live streaming transcription via WebSocket","JupyterLab for batch processing scripts","SRT/VTT subtitle export"],"env_var_fields":[{"key":"WHISPER_MODEL","label":"Model","options":[{"label":"Large V3 Turbo — best speed/accuracy ratio, ~2.5 GB VRAM (recommended)","value":"deepdml/faster-whisper-large-v3-turbo-ct2"},{"label":"Distil Large V3 — fast, near-large quality, ~2.4 GB VRAM","value":"Systran/faster-distil-whisper-large-v3"},{"label":"Large V3 — highest accuracy, ~4.5 GB VRAM","value":"Systran/faster-whisper-large-v3"},{"label":"Large V2 — proven quality, ~4.5 GB VRAM","value":"Systran/faster-whisper-large-v2"},{"label":"Medium — balanced, ~2.5 GB VRAM","value":"Systran/faster-whisper-medium"},{"label":"Small — lightweight, ~0.8 GB VRAM","value":"Systran/faster-whisper-small"}],"required":true,"placeholder":"deepdml/faster-whisper-large-v3-turbo-ct2"}],"show_gpu_selector":true,"expanded_description":"GPU-accelerated speech-to-text and text-to-speech powered by Faster Whisper and Kokoro TTS. Up to 4x faster than OpenAI's Whisper with the same accuracy. Drag and drop audio or video files into the web UI for instant transcription, generate natural English speech with Kokoro TTS, or use the OpenAI-compatible API for batch processing. Supports 99 languages, subtitle generation (SRT/VTT), and live streaming transcription via WebSocket.","optional_env_var_fields":[{"key":"WHISPER_COMPUTE_TYPE","label":"Precision","required":false,"help_link":{"url":"https://speaches.ai/configuration/","text":"float16 (GPU default) or int8 (lower VRAM)"},"sensitive":false,"placeholder":"float16"}]},"addons":{"available":[{"id":"jupyter","name":"JupyterLab","description":"Browser-based Python IDE with GPU monitoring","port":8888,"access_label":"JupyterLab","access_method":"jupyter"},{"id":"code-server","name":"VS Code Server","description":"Full VS Code editor in your browser","port":8443,"access_label":"VS Code","access_method":"code_server"}],"defaults":["jupyter"]}},{"id":"00000000-0000-0000-0000-000000000010","name":"Jupyter + PyTorch","slug":"jupyter-gpu","description":"GPU-accelerated ML environment with JupyterLab, PyTorch, and CUDA. Your blank canvas for training, fine-tuning, and experimentation.","docker_image":"ghcr.io/benpoliquin/templates/jupyter-gpu@sha256:5e4b36d0cce84a5acfae2080c0762eb4e39296171f9e1767505908e0d1f9c817","requires_gpu":true,"min_gpu":"rtx-a6000","min_vram_gb":24,"estimated_setup_seconds":120,"category":"development","source_repo":"","billing_model":"per_minute","proxy_port":8888,"monthly_price_cents":null,"volume_gb":50,"base_image_type":"custom","is_active":true,"datasets":[],"gpu_options":[{"id":"36e09bb9-b49d-43b1-b9e3-1eea1ede5f3a","template_id":"00000000-0000-0000-0000-000000000010","gpu_type":"rtx-a6000","display_name":"RTX A6000","sell_price_per_hr_cents":97,"is_default":true,"sort_order":0,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":200,"available":true,"stock_level":"low","providers":[{"provider":"runpod","sell_price_per_hr_cents":97,"available":1,"stock_level":"low","region":"","cloud_tier":"SECURE"}],"vram_gb":48,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"d495a024-09ac-4a07-a84d-b48b876a9457","template_id":"00000000-0000-0000-0000-000000000010","gpu_type":"rtx-4090","display_name":"RTX 4090","sell_price_per_hr_cents":48,"is_default":false,"sort_order":1,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":150,"available":true,"stock_level":"medium","providers":[{"provider":"runpod","sell_price_per_hr_cents":48,"available":1,"stock_level":"low","region":"","cloud_tier":"COMMUNITY"},{"provider":"vastai","sell_price_per_hr_cents":57,"available":1,"stock_level":"medium","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":61,"available":1,"stock_level":"medium","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"Pennsylvania, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"British Columbia, CA","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":85,"available":1,"stock_level":"low","region":"Minnesota, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":12,"max_vcpus":256,"min_ram_gb":31,"max_ram_gb":126},{"id":"c2bcc4e9-ef06-4050-926d-afc08d414415","template_id":"00000000-0000-0000-0000-000000000010","gpu_type":"a100-80gb-sxm","display_name":"A100 80GB SXM","sell_price_per_hr_cents":176,"is_default":false,"sort_order":2,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":500,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":176,"available":1,"stock_level":"low","region":"Montana, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":187,"available":1,"stock_level":"low","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":195,"available":1,"stock_level":"low","region":"North Carolina, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":307,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":80,"min_vcpus":128,"max_vcpus":208,"min_ram_gb":126,"max_ram_gb":126},{"id":"1d671329-421d-42ed-aca4-5f9bf8ba0ab3","template_id":"00000000-0000-0000-0000-000000000010","gpu_type":"h100-sxm","display_name":"H100 SXM","sell_price_per_hr_cents":615,"is_default":false,"sort_order":3,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":500,"available":true,"stock_level":"high","providers":[{"provider":"runpod","sell_price_per_hr_cents":615,"available":1,"stock_level":"high","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"}],"vram_gb":80,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"2097da34-48c4-4083-95a3-125663a4cbb8","template_id":"00000000-0000-0000-0000-000000000010","gpu_type":"2xa100-80gb-sxm","display_name":"2xA100 80GB SXM","sell_price_per_hr_cents":374,"is_default":false,"sort_order":4,"compute_type":"gpu","gpu_count":2,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":374,"available":1,"stock_level":"low","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":389,"available":1,"stock_level":"low","region":"North Carolina, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":614,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":160,"min_vcpus":128,"max_vcpus":256,"min_ram_gb":252,"max_ram_gb":252},{"id":"76c81633-ec0f-4e81-bf9c-326c2c9820ff","template_id":"00000000-0000-0000-0000-000000000010","gpu_type":"2xh100-sxm","display_name":"2xH100 SXM","sell_price_per_hr_cents":486,"is_default":false,"sort_order":5,"compute_type":"gpu","gpu_count":2,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":486,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":1230,"available":1,"stock_level":"high","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"}],"vram_gb":160,"min_vcpus":256,"max_vcpus":256,"min_ram_gb":252,"max_ram_gb":252},{"id":"1aa8d9e6-4686-4d94-8c91-698ed1518206","template_id":"00000000-0000-0000-0000-000000000010","gpu_type":"4xa100-80gb-sxm","display_name":"4xA100 80GB SXM","sell_price_per_hr_cents":1227,"is_default":false,"sort_order":6,"compute_type":"gpu","gpu_count":4,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"runpod","sell_price_per_hr_cents":1227,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":320,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"a14e1698-5d2c-4c1d-b044-5bb1e6e27002","template_id":"00000000-0000-0000-0000-000000000010","gpu_type":"4xh100-sxm","display_name":"4xH100 SXM","sell_price_per_hr_cents":971,"is_default":false,"sort_order":7,"compute_type":"gpu","gpu_count":4,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":971,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":2459,"available":1,"stock_level":"high","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"}],"vram_gb":320,"min_vcpus":256,"max_vcpus":256,"min_ram_gb":504,"max_ram_gb":504},{"id":"1ba5ccc4-56f0-4363-9df8-051ad0c2d710","template_id":"00000000-0000-0000-0000-000000000010","gpu_type":"8xa100-80gb-sxm","display_name":"8xA100 80GB SXM","sell_price_per_hr_cents":2453,"is_default":false,"sort_order":8,"compute_type":"gpu","gpu_count":8,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"low","providers":[{"provider":"runpod","sell_price_per_hr_cents":2453,"available":1,"stock_level":"low","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":640,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"813dc8e2-ae49-4d7c-8d16-f0db59bf9f8f","template_id":"00000000-0000-0000-0000-000000000010","gpu_type":"8xh100-sxm","display_name":"8xH100 SXM","sell_price_per_hr_cents":4917,"is_default":false,"sort_order":9,"compute_type":"gpu","gpu_count":8,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"low","providers":[{"provider":"runpod","sell_price_per_hr_cents":4917,"available":1,"stock_level":"low","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"}],"vram_gb":640,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0}],"frontend_config":{"gpu_badges":{"rtx-4090":"Best price-to-performance"},"gpu_guidance":{"h100-sxm":"80 GB HBM3 — 2-3x faster than A100. For time-critical training runs.","rtx-4090":"24 GB VRAM — fastest consumer GPU. Best throughput for workloads that fit in 24 GB.","rtx-a6000":"48 GB VRAM — best VRAM per dollar. Fine-tune 7B-13B models, plenty of headroom.","2xh100-sxm":"160 GB VRAM — 70B+ in FP16 with NVLink. For large models and fast distributed training.","4xh100-sxm":"320 GB VRAM — full 70B training or frontier model experiments with tensor parallelism.","8xh100-sxm":"640 GB VRAM — full DGX-class node. For frontier model training and research.","a100-80gb-sxm":"80 GB HBM2e — datacenter training workhorse. 70B quantized or 30B+ in FP16.","2xa100-80gb-sxm":"160 GB VRAM — train 70B models in FP16 or run large-scale distributed experiments.","4xa100-80gb-sxm":"320 GB VRAM — large-scale distributed training. Full 70B in FP16 with data parallelism.","8xa100-80gb-sxm":"640 GB VRAM — maximum A100 scale. Train 100B+ models or massive data-parallel jobs."},"what_you_get":["JupyterLab with GPU access","PyTorch + CUDA pre-installed","Persistent /workspace storage","Full terminal and SSH access"],"env_var_fields":[],"contact_for_more":"Need custom configurations or dedicated clusters?","show_gpu_selector":true,"expanded_description":"A GPU-accelerated development environment with JupyterLab, PyTorch, and full CUDA toolkit. Clone any repo, pip install any package, train any model. Your blank canvas for ML experimentation.","optional_env_var_fields":[]},"addons":{"available":[{"id":"code-server","name":"VS Code Server","description":"Full VS Code editor in your browser","port":8443,"access_label":"VS Code","access_method":"code_server"}],"defaults":[]}},{"id":"00000000-0000-0000-0000-000000000031","name":"Kokoro TTS","slug":"kokoro-tts","description":"OpenAI-compatible text-to-speech on GPU. 54 voices, 8 languages, near-instant start.","docker_image":"ghcr.io/benpoliquin/templates/kokoro-tts@sha256:8f50535946e6993db4f125a5ef14b1a3970b5f878cb68c95013b9bc7f90ea240","requires_gpu":true,"min_gpu":"rtx-3090","min_vram_gb":8,"estimated_setup_seconds":60,"category":"audio","source_repo":"https://github.com/remsky/Kokoro-FastAPI","billing_model":"per_minute","proxy_port":8880,"monthly_price_cents":null,"volume_gb":20,"base_image_type":"custom","is_active":true,"datasets":[],"gpu_options":[{"id":"75563578-9373-4307-9f78-b63aef93bcdc","template_id":"00000000-0000-0000-0000-000000000031","gpu_type":"rtx-3090","display_name":"RTX 3090","sell_price_per_hr_cents":25,"is_default":true,"sort_order":0,"compute_type":"gpu","gpu_count":1,"default_volume_gb":20,"max_volume_gb":100,"available":true,"stock_level":"medium","providers":[{"provider":"vastai","sell_price_per_hr_cents":25,"available":1,"stock_level":"medium","region":"United States, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":29,"available":1,"stock_level":"medium","region":"Nevada, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":36,"available":1,"stock_level":"low","region":"Utah, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":66,"available":1,"stock_level":"low","region":"Ohio, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":28,"max_vcpus":96,"min_ram_gb":15,"max_ram_gb":126},{"id":"e67900ed-044a-434a-885e-fe36fe8238e2","template_id":"00000000-0000-0000-0000-000000000031","gpu_type":"rtx-4090","display_name":"RTX 4090","sell_price_per_hr_cents":48,"is_default":false,"sort_order":1,"compute_type":"gpu","gpu_count":1,"default_volume_gb":20,"max_volume_gb":100,"available":true,"stock_level":"medium","providers":[{"provider":"runpod","sell_price_per_hr_cents":48,"available":1,"stock_level":"low","region":"","cloud_tier":"COMMUNITY"},{"provider":"vastai","sell_price_per_hr_cents":57,"available":1,"stock_level":"medium","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":61,"available":1,"stock_level":"medium","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"Pennsylvania, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"British Columbia, CA","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":85,"available":1,"stock_level":"low","region":"Minnesota, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":12,"max_vcpus":256,"min_ram_gb":31,"max_ram_gb":126}],"frontend_config":{"gpu_guidance":{"rtx-3090":"24 GB VRAM — far more than needed (~3 GB); great value.","rtx-4090":"Fastest single-GPU synthesis (100x+ realtime)."},"what_you_get":["OpenAI-compatible /v1/audio/speech endpoint","54 voices across 8 languages","Built-in web UI + Swagger docs"],"env_var_fields":[],"show_gpu_selector":true,"expanded_description":"Kokoro-82M (Apache 2.0) served with an OpenAI-compatible /v1/audio/speech endpoint plus a built-in web UI at /web. Point any OpenAI TTS client at your instance URL. Model is baked in — starts in seconds.","optional_env_var_fields":[]}},{"id":"00000000-0000-0000-0000-000000000032","name":"LLaMA-Factory","slug":"llama-factory","description":"GUI fine-tuning for 100+ LLMs. LoRA/QLoRA from your browser with LlamaBoard — no code required.","docker_image":"ghcr.io/benpoliquin/templates/llama-factory@sha256:6f8eee1709feb349d66972381627828583817dbba9df40f52956bc0b79960e3d","requires_gpu":true,"min_gpu":"rtx-3090","min_vram_gb":16,"estimated_setup_seconds":120,"category":"development","source_repo":"https://github.com/hiyouga/LLaMA-Factory","billing_model":"per_minute","proxy_port":7860,"monthly_price_cents":null,"volume_gb":50,"base_image_type":"custom","is_active":true,"datasets":[],"gpu_options":[{"id":"2ccda822-418f-4228-8898-91d192935284","template_id":"00000000-0000-0000-0000-000000000032","gpu_type":"rtx-3090","display_name":"RTX 3090","sell_price_per_hr_cents":25,"is_default":true,"sort_order":0,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":200,"available":true,"stock_level":"medium","providers":[{"provider":"vastai","sell_price_per_hr_cents":25,"available":1,"stock_level":"medium","region":"United States, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":29,"available":1,"stock_level":"medium","region":"Nevada, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":36,"available":1,"stock_level":"low","region":"Utah, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":66,"available":1,"stock_level":"low","region":"Ohio, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":28,"max_vcpus":96,"min_ram_gb":15,"max_ram_gb":126},{"id":"dd1e71d5-557d-42dd-b9b7-f30e5fc0aa95","template_id":"00000000-0000-0000-0000-000000000032","gpu_type":"rtx-a6000","display_name":"RTX A6000","sell_price_per_hr_cents":57,"is_default":false,"sort_order":1,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":200,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":57,"available":1,"stock_level":"low","region":"Kansas, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":97,"available":1,"stock_level":"low","region":"","cloud_tier":"SECURE"},{"provider":"lambda_labs","sell_price_per_hr_cents":153,"available":1,"stock_level":"high","region":"us-south-2","cloud_tier":"ON_DEMAND"}],"vram_gb":48,"min_vcpus":14,"max_vcpus":64,"min_ram_gb":63,"max_ram_gb":100},{"id":"ac4a4e04-a77f-456f-bb0f-0cd961e1f91c","template_id":"00000000-0000-0000-0000-000000000032","gpu_type":"rtx-4090","display_name":"RTX 4090","sell_price_per_hr_cents":48,"is_default":false,"sort_order":2,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":200,"available":true,"stock_level":"medium","providers":[{"provider":"runpod","sell_price_per_hr_cents":48,"available":1,"stock_level":"low","region":"","cloud_tier":"COMMUNITY"},{"provider":"vastai","sell_price_per_hr_cents":57,"available":1,"stock_level":"medium","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":61,"available":1,"stock_level":"medium","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"Pennsylvania, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"British Columbia, CA","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":85,"available":1,"stock_level":"low","region":"Minnesota, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":12,"max_vcpus":256,"min_ram_gb":31,"max_ram_gb":126},{"id":"3cb8285f-a18d-4039-b12e-d68556e9f458","template_id":"00000000-0000-0000-0000-000000000032","gpu_type":"a100-80gb-sxm","display_name":"A100 80GB SXM","sell_price_per_hr_cents":176,"is_default":false,"sort_order":3,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":500,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":176,"available":1,"stock_level":"low","region":"Montana, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":187,"available":1,"stock_level":"low","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":195,"available":1,"stock_level":"low","region":"North Carolina, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":307,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":80,"min_vcpus":128,"max_vcpus":208,"min_ram_gb":126,"max_ram_gb":126}],"frontend_config":{"gpu_guidance":{"rtx-3090":"24 GB VRAM — QLoRA up to ~30B models; best value to start.","rtx-4090":"24 GB VRAM — fastest consumer card for 7B-14B fine-tunes.","rtx-a6000":"48 GB VRAM — QLoRA up to 70B, LoRA-fp16 mid-size models.","a100-80gb-sxm":"80 GB VRAM — full LoRA on large models, fastest training."},"what_you_get":["LlamaBoard web GUI — fine-tune without writing code","LoRA / QLoRA for 100+ models (Llama, Qwen, Mistral, Gemma)","Bundled sample datasets, custom dataset upload","Checkpoints persist in /workspace/output","JupyterLab for scripted runs and evaluation"],"env_var_fields":[],"show_gpu_selector":true,"expanded_description":"The most-adopted open fine-tuning framework (70k+ stars, Apache 2.0) with the LlamaBoard web GUI. Pick a model, pick a dataset (bundled samples included), and run LoRA or QLoRA fine-tunes from your browser — checkpoints land in persistent storage and survive restarts. 7B QLoRA fits in 6 GB VRAM; 70B QLoRA in 48 GB.","optional_env_var_fields":[{"key":"HF_TOKEN","label":"Hugging Face token","required":false,"help_link":{"url":"https://huggingface.co/settings/tokens","text":"Needed for gated models (Llama, Gemma)"},"sensitive":true,"placeholder":"hf_..."}]},"addons":{"available":[{"id":"jupyter","name":"JupyterLab","description":"Browser-based Python IDE with GPU monitoring","port":8888,"access_label":"JupyterLab","access_method":"jupyter"},{"id":"code-server","name":"VS Code Server","description":"Full VS Code editor in your browser","port":8443,"access_label":"VS Code","access_method":"code_server"}],"defaults":["jupyter"]}},{"id":"00000000-0000-0000-0000-000000000011","name":"Ollama","slug":"ollama","description":"Run open-source LLMs on GPU with a ChatGPT-like UI. Pick your model, start chatting in minutes.","docker_image":"ghcr.io/benpoliquin/templates/ollama@sha256:3a173b7dfc5fdcb44ec0344e514571b22f6f2fd31e154b0977351934f4108c3a","requires_gpu":true,"min_gpu":"rtx-3090","min_vram_gb":24,"estimated_setup_seconds":180,"category":"language-models","source_repo":"https://github.com/ollama/ollama","billing_model":"per_minute","proxy_port":8080,"monthly_price_cents":null,"volume_gb":20,"base_image_type":"custom","is_active":true,"datasets":[],"gpu_options":[{"id":"bf28e437-29bf-4d25-a24a-e21fe855beca","template_id":"00000000-0000-0000-0000-000000000011","gpu_type":"rtx-3090","display_name":"RTX 3090","sell_price_per_hr_cents":25,"is_default":true,"sort_order":0,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":150,"available":true,"stock_level":"medium","providers":[{"provider":"vastai","sell_price_per_hr_cents":25,"available":1,"stock_level":"medium","region":"United States, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":29,"available":1,"stock_level":"medium","region":"Nevada, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":36,"available":1,"stock_level":"low","region":"Utah, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":66,"available":1,"stock_level":"low","region":"Ohio, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":28,"max_vcpus":96,"min_ram_gb":15,"max_ram_gb":126},{"id":"7432ef6a-23c1-44de-a0e8-672cefbf628b","template_id":"00000000-0000-0000-0000-000000000011","gpu_type":"rtx-a6000","display_name":"RTX A6000","sell_price_per_hr_cents":57,"is_default":false,"sort_order":1,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":200,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":57,"available":1,"stock_level":"low","region":"Kansas, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":97,"available":1,"stock_level":"low","region":"","cloud_tier":"SECURE"},{"provider":"lambda_labs","sell_price_per_hr_cents":153,"available":1,"stock_level":"high","region":"us-south-2","cloud_tier":"ON_DEMAND"}],"vram_gb":48,"min_vcpus":14,"max_vcpus":64,"min_ram_gb":63,"max_ram_gb":100},{"id":"b0d4bd38-c980-43ad-8469-eb17d85496dc","template_id":"00000000-0000-0000-0000-000000000011","gpu_type":"rtx-4090","display_name":"RTX 4090","sell_price_per_hr_cents":48,"is_default":false,"sort_order":2,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":150,"available":true,"stock_level":"medium","providers":[{"provider":"runpod","sell_price_per_hr_cents":48,"available":1,"stock_level":"low","region":"","cloud_tier":"COMMUNITY"},{"provider":"vastai","sell_price_per_hr_cents":57,"available":1,"stock_level":"medium","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":61,"available":1,"stock_level":"medium","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"Pennsylvania, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"British Columbia, CA","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":85,"available":1,"stock_level":"low","region":"Minnesota, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":12,"max_vcpus":256,"min_ram_gb":31,"max_ram_gb":126},{"id":"6e95511d-6f2e-490d-868a-19e99b1a8456","template_id":"00000000-0000-0000-0000-000000000011","gpu_type":"a100-80gb-sxm","display_name":"A100 80GB SXM","sell_price_per_hr_cents":176,"is_default":false,"sort_order":3,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":500,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":176,"available":1,"stock_level":"low","region":"Montana, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":187,"available":1,"stock_level":"low","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":195,"available":1,"stock_level":"low","region":"North Carolina, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":307,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":80,"min_vcpus":128,"max_vcpus":208,"min_ram_gb":126,"max_ram_gb":126},{"id":"618d417b-7fb1-42f8-8d25-fdfa0a330aad","template_id":"00000000-0000-0000-0000-000000000011","gpu_type":"2xh100-sxm","display_name":"2xH100 SXM","sell_price_per_hr_cents":486,"is_default":false,"sort_order":4,"compute_type":"gpu","gpu_count":2,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":486,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":1230,"available":1,"stock_level":"high","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"},{"provider":"lambda_labs","sell_price_per_hr_cents":1342,"available":0,"stock_level":"none","region":"","cloud_tier":"ON_DEMAND"}],"vram_gb":160,"min_vcpus":52,"max_vcpus":256,"min_ram_gb":252,"max_ram_gb":450},{"id":"837e1c5a-6e31-46d3-b9d1-750ca84fe532","template_id":"00000000-0000-0000-0000-000000000011","gpu_type":"4xh100-sxm","display_name":"4xH100 SXM","sell_price_per_hr_cents":971,"is_default":false,"sort_order":5,"compute_type":"gpu","gpu_count":4,"default_volume_gb":300,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":971,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":2459,"available":1,"stock_level":"high","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"},{"provider":"lambda_labs","sell_price_per_hr_cents":2627,"available":0,"stock_level":"none","region":"","cloud_tier":"ON_DEMAND"}],"vram_gb":320,"min_vcpus":104,"max_vcpus":256,"min_ram_gb":504,"max_ram_gb":900},{"id":"19988412-fe4e-4a7b-9819-51a9de3c6b1b","template_id":"00000000-0000-0000-0000-000000000011","gpu_type":"8xh100-sxm","display_name":"8xH100 SXM","sell_price_per_hr_cents":4917,"is_default":false,"sort_order":6,"compute_type":"gpu","gpu_count":8,"default_volume_gb":500,"max_volume_gb":2000,"available":true,"stock_level":"low","providers":[{"provider":"runpod","sell_price_per_hr_cents":4917,"available":1,"stock_level":"low","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"},{"provider":"lambda_labs","sell_price_per_hr_cents":5141,"available":0,"stock_level":"none","region":"","cloud_tier":"ON_DEMAND"}],"vram_gb":640,"min_vcpus":208,"max_vcpus":208,"min_ram_gb":1800,"max_ram_gb":1800}],"frontend_config":{"gpu_guidance":{"rtx-3090":"24 GB VRAM — runs all 4B-14B models smoothly (Qwen 3 8B, Llama 3.1 8B, Phi-4 Mini, gpt-oss 20B). Best value for chat.","rtx-4090":"24 GB VRAM — fastest 24 GB card. Snappiest responses for 7B-13B models.","rtx-a6000":"48 GB VRAM — unlocks 24B-32B models (Qwen 3 32B, Gemma 3 27B, Qwen3-Coder 30B, Magistral 24B). Best VRAM per dollar.","2xh100-sxm":"160 GB VRAM — Llama 4 Scout MoE (109B) or frontier 70B with long context.","4xh100-sxm":"320 GB VRAM — Qwen 3 235B-A22B MoE or dense frontier models with tensor parallelism.","8xh100-sxm":"640 GB VRAM — DeepSeek V3 671B or Llama 4 Maverick (400B MoE) at full scale.","a100-80gb-sxm":"80 GB VRAM — single-GPU home for 70B models (Llama 3.3, DeepSeek R1) and gpt-oss 120B MoE."},"what_you_get":["ChatGPT-like web UI (Open WebUI)","GPU-accelerated LLM inference","Pull additional models anytime"],"env_var_fields":[{"key":"OLLAMA_MODEL","label":"Model","options":[{"label":"Qwen 3 8B — 5.2 GB download, ~6 GB VRAM (thinking mode)","value":"qwen3:8b"},{"label":"Llama 3.1 8B — 4.9 GB download, ~6 GB VRAM","value":"llama3.1:8b"},{"label":"Phi-4 Mini 3.8B — 2.5 GB download, ~4 GB VRAM","value":"phi4-mini:3.8b"},{"label":"Gemma 3 4B — 3.3 GB download, ~4 GB VRAM (vision, 128K ctx)","value":"gemma3:4b"},{"label":"Qwen3-Coder 30B MoE — 19 GB download, 3.3B active, 256K ctx","value":"qwen3-coder:30b","recommended_gpu":"rtx-a6000"},{"label":"Qwen 2.5 Coder 32B — 20 GB download, ~22 GB VRAM","value":"qwen2.5-coder:32b","recommended_gpu":"rtx-a6000"},{"label":"DeepSeek R1 8B — 5.2 GB download, ~6 GB VRAM (reasoning)","value":"deepseek-r1:8b"},{"label":"gpt-oss 20B MoE — 14 GB download, fits 16 GB VRAM (reasoning/agents)","value":"gpt-oss:20b"},{"label":"DeepSeek R1 32B — 20 GB download, ~20 GB VRAM (reasoning)","value":"deepseek-r1:32b","recommended_gpu":"rtx-a6000"},{"label":"Magistral 24B — 14 GB download, ~16 GB VRAM (reasoning)","value":"magistral:24b","recommended_gpu":"rtx-a6000"},{"label":"Qwen 3 32B — 20 GB download, ~20 GB VRAM","value":"qwen3:32b","recommended_gpu":"rtx-a6000"},{"label":"Gemma 3 27B — 17 GB download, ~17 GB VRAM (vision, 128K ctx)","value":"gemma3:27b","recommended_gpu":"rtx-a6000"},{"label":"Gemma 4 26B MoE — 18 GB download, 4B active, 256K ctx","value":"gemma4:26b","recommended_gpu":"rtx-a6000"},{"label":"Mistral Small 3.2 24B — 15 GB download, ~16 GB VRAM (vision)","value":"mistral-small3.2:24b","recommended_gpu":"rtx-a6000"},{"label":"Llama 3.3 70B — 43 GB download, ~45 GB VRAM","value":"llama3.3:70b","recommended_gpu":"a100-80gb-sxm"},{"label":"DeepSeek R1 70B — 43 GB download, ~45 GB VRAM (reasoning)","value":"deepseek-r1:70b","recommended_gpu":"a100-80gb-sxm"},{"label":"gpt-oss 120B MoE — 65 GB download, fits single 80 GB (reasoning)","value":"gpt-oss:120b","recommended_gpu":"a100-80gb-sxm"},{"label":"Llama 4 Scout 17B×16E MoE — 67 GB download, 109B total, 10M ctx","value":"llama4:scout","recommended_gpu":"2xh100-sxm"},{"label":"Qwen 3 235B-A22B MoE — 142 GB download, 22B active, 256K ctx","value":"qwen3:235b","recommended_gpu":"4xh100-sxm"},{"label":"Llama 4 Maverick 17B×128E MoE — 245 GB download, 400B total","value":"llama4:maverick","recommended_gpu":"8xh100-sxm"},{"label":"DeepSeek V3 671B MoE — 404 GB download, 37B active (Q4)","value":"deepseek-v3:671b","recommended_gpu":"8xh100-sxm"}],"required":true,"placeholder":"qwen3:8b","allow_custom":false}],"show_gpu_selector":true,"expanded_description":"Run open-source LLMs on GPU hardware with Ollama. Includes Open WebUI for a ChatGPT-like chat interface and JupyterLab for programmatic access via the Ollama Python SDK. Pick your model, pick your GPU, start chatting.","optional_env_var_fields":[]},"addons":{"available":[{"id":"jupyter","name":"JupyterLab","description":"Browser-based Python IDE with GPU monitoring","port":8888,"access_label":"JupyterLab","access_method":"jupyter"},{"id":"code-server","name":"VS Code Server","description":"Full VS Code editor in your browser","port":8443,"access_label":"VS Code","access_method":"code_server"}],"defaults":["jupyter"]}},{"id":"00000000-0000-0000-0000-000000000015","name":"Unsloth","slug":"unsloth","description":"LLM and decision-model fine-tuning with LoRA and QLoRA. Unsloth Studio no-code UI plus JupyterLab with a pre-loaded quickstart notebook.","docker_image":"ghcr.io/benpoliquin/templates/unsloth@sha256:90c9c8c11da704f7cc33decde3bec78697b56579fa9c9f37003d2038282214c8","requires_gpu":true,"min_gpu":"rtx-a6000","min_vram_gb":16,"estimated_setup_seconds":600,"category":"language-models","source_repo":"https://github.com/unslothai/unsloth","billing_model":"per_minute","proxy_port":8000,"monthly_price_cents":null,"volume_gb":50,"base_image_type":"custom","is_active":true,"datasets":[],"gpu_options":[{"id":"460be02f-cca5-44c5-881b-f81ec73b1833","template_id":"00000000-0000-0000-0000-000000000015","gpu_type":"rtx-a6000","display_name":"RTX A6000","sell_price_per_hr_cents":57,"is_default":true,"sort_order":0,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":200,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":57,"available":1,"stock_level":"low","region":"Kansas, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":97,"available":1,"stock_level":"low","region":"","cloud_tier":"SECURE"},{"provider":"lambda_labs","sell_price_per_hr_cents":153,"available":1,"stock_level":"high","region":"us-south-2","cloud_tier":"ON_DEMAND"}],"vram_gb":48,"min_vcpus":14,"max_vcpus":64,"min_ram_gb":63,"max_ram_gb":100},{"id":"dceedc2c-8c56-428d-8f80-98565f2b84bd","template_id":"00000000-0000-0000-0000-000000000015","gpu_type":"rtx-4090","display_name":"RTX 4090","sell_price_per_hr_cents":48,"is_default":false,"sort_order":1,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":200,"available":true,"stock_level":"medium","providers":[{"provider":"runpod","sell_price_per_hr_cents":48,"available":1,"stock_level":"low","region":"","cloud_tier":"COMMUNITY"},{"provider":"vastai","sell_price_per_hr_cents":57,"available":1,"stock_level":"medium","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":61,"available":1,"stock_level":"medium","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"Pennsylvania, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"British Columbia, CA","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":85,"available":1,"stock_level":"low","region":"Minnesota, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":12,"max_vcpus":256,"min_ram_gb":31,"max_ram_gb":126},{"id":"2179667d-fc0d-4fb2-a6d1-f1da67b65de3","template_id":"00000000-0000-0000-0000-000000000015","gpu_type":"a100-80gb-sxm","display_name":"A100 80GB SXM","sell_price_per_hr_cents":176,"is_default":false,"sort_order":2,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":500,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":176,"available":1,"stock_level":"low","region":"Montana, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":187,"available":1,"stock_level":"low","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":195,"available":1,"stock_level":"low","region":"North Carolina, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":307,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":80,"min_vcpus":128,"max_vcpus":208,"min_ram_gb":126,"max_ram_gb":126},{"id":"77bacc78-4283-4b70-a029-628d0b3e4772","template_id":"00000000-0000-0000-0000-000000000015","gpu_type":"h100-sxm","display_name":"H100 SXM","sell_price_per_hr_cents":615,"is_default":false,"sort_order":3,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":500,"available":true,"stock_level":"high","providers":[{"provider":"runpod","sell_price_per_hr_cents":615,"available":1,"stock_level":"high","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"},{"provider":"lambda_labs","sell_price_per_hr_cents":685,"available":0,"stock_level":"none","region":"","cloud_tier":"ON_DEMAND"}],"vram_gb":80,"min_vcpus":26,"max_vcpus":26,"min_ram_gb":225,"max_ram_gb":225},{"id":"8521270a-4f28-4ee5-9869-ef28b6177048","template_id":"00000000-0000-0000-0000-000000000015","gpu_type":"2xa100-80gb-sxm","display_name":"2xA100 80GB SXM","sell_price_per_hr_cents":374,"is_default":false,"sort_order":4,"compute_type":"gpu","gpu_count":2,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":374,"available":1,"stock_level":"low","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":389,"available":1,"stock_level":"low","region":"North Carolina, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":614,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":160,"min_vcpus":128,"max_vcpus":256,"min_ram_gb":252,"max_ram_gb":252},{"id":"374c89eb-8145-4544-ad67-8ff5dbc6fe10","template_id":"00000000-0000-0000-0000-000000000015","gpu_type":"2xh100-sxm","display_name":"2xH100 SXM","sell_price_per_hr_cents":486,"is_default":false,"sort_order":5,"compute_type":"gpu","gpu_count":2,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":486,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":1230,"available":1,"stock_level":"high","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"},{"provider":"lambda_labs","sell_price_per_hr_cents":1342,"available":0,"stock_level":"none","region":"","cloud_tier":"ON_DEMAND"}],"vram_gb":160,"min_vcpus":52,"max_vcpus":256,"min_ram_gb":252,"max_ram_gb":450},{"id":"32db9ad1-f28b-402e-bc67-569570078d47","template_id":"00000000-0000-0000-0000-000000000015","gpu_type":"4xa100-80gb-sxm","display_name":"4xA100 80GB SXM","sell_price_per_hr_cents":1227,"is_default":false,"sort_order":6,"compute_type":"gpu","gpu_count":4,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"runpod","sell_price_per_hr_cents":1227,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":320,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"73ca6127-3199-46e4-aaf6-679303d9ca55","template_id":"00000000-0000-0000-0000-000000000015","gpu_type":"4xh100-sxm","display_name":"4xH100 SXM","sell_price_per_hr_cents":971,"is_default":false,"sort_order":7,"compute_type":"gpu","gpu_count":4,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":971,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":2459,"available":1,"stock_level":"high","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"},{"provider":"lambda_labs","sell_price_per_hr_cents":2627,"available":0,"stock_level":"none","region":"","cloud_tier":"ON_DEMAND"}],"vram_gb":320,"min_vcpus":104,"max_vcpus":256,"min_ram_gb":504,"max_ram_gb":900},{"id":"7aef02bc-28c4-41cf-a76a-2c1dbfeb76c1","template_id":"00000000-0000-0000-0000-000000000015","gpu_type":"8xa100-80gb-sxm","display_name":"8xA100 80GB SXM","sell_price_per_hr_cents":2453,"is_default":false,"sort_order":8,"compute_type":"gpu","gpu_count":8,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"low","providers":[{"provider":"runpod","sell_price_per_hr_cents":2453,"available":1,"stock_level":"low","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"},{"provider":"lambda_labs","sell_price_per_hr_cents":3573,"available":0,"stock_level":"none","region":"","cloud_tier":"ON_DEMAND"}],"vram_gb":640,"min_vcpus":240,"max_vcpus":240,"min_ram_gb":1800,"max_ram_gb":1800},{"id":"daee5523-b386-4343-b40e-af53887cbe74","template_id":"00000000-0000-0000-0000-000000000015","gpu_type":"8xh100-sxm","display_name":"8xH100 SXM","sell_price_per_hr_cents":4917,"is_default":false,"sort_order":9,"compute_type":"gpu","gpu_count":8,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"low","providers":[{"provider":"runpod","sell_price_per_hr_cents":4917,"available":1,"stock_level":"low","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"},{"provider":"lambda_labs","sell_price_per_hr_cents":5141,"available":0,"stock_level":"none","region":"","cloud_tier":"ON_DEMAND"}],"vram_gb":640,"min_vcpus":208,"max_vcpus":208,"min_ram_gb":1800,"max_ram_gb":1800}],"frontend_config":{"gpu_badges":{"rtx-4090":"Best price-to-performance"},"gpu_guidance":{"h100-sxm":"80 GB VRAM — 2-3x faster than A100 with FP8 Transformer Engine. For time-critical runs.","rtx-4090":"24 GB VRAM — fastest 24 GB card. Ideal for 7B-8B QLoRA with maximum throughput.","rtx-a6000":"48 GB VRAM — best VRAM per dollar. Comfortably fine-tunes 7B-13B with QLoRA, even 27B.","2xh100-sxm":"160 GB VRAM — fastest multi-GPU fine-tuning. 70B full LoRA with NVLink bandwidth.","4xh100-sxm":"320 GB VRAM — fastest 4-GPU fine-tuning. Full LoRA on 70B+ with FP8 Transformer Engine.","8xh100-sxm":"640 GB VRAM — DGX-class node. For frontier-scale fine-tuning.","a100-80gb-sxm":"80 GB VRAM — fine-tune 70B with QLoRA or 13B with full LoRA. HBM2e bandwidth.","2xa100-80gb-sxm":"160 GB VRAM — fine-tune 70B+ models with full LoRA or large QLoRA runs.","4xa100-80gb-sxm":"320 GB VRAM — fine-tune 100B+ models or run large data-parallel training jobs.","8xa100-80gb-sxm":"640 GB VRAM — maximum A100 scale for the largest fine-tuning runs."},"what_you_get":["Unsloth Studio — no-code fine-tuning UI","LoRA/QLoRA for supported LLMs and decision models","JupyterLab with pre-loaded quickstart notebook","Export to GGUF, merged weights, or HuggingFace Hub","Persistent /workspace for models, datasets, and checkpoints"],"env_var_fields":[],"contact_for_more":"Need custom configurations for large-scale fine-tuning?","show_gpu_selector":true,"expanded_description":"Fine-tune supported LLMs and decision models using Unsloth. Includes Unsloth Studio, a no-code web UI for fine-tuning, dataset creation, model export, and chat. Supports LoRA and QLoRA on consumer and datacenter GPUs. Also comes with JupyterLab and a pre-loaded quickstart notebook that walks you through loading a model, configuring LoRA adapters, training with SFTTrainer, and saving your fine-tuned model. Export supported language models to GGUF for Ollama/llama.cpp, or push to Hugging Face Hub. Initial setup can take about 10 minutes while the image downloads and Studio initializes.","optional_env_var_fields":[{"key":"HF_TOKEN","label":"HuggingFace Token","required":false,"help_link":{"url":"https://huggingface.co/settings/tokens","text":"Get a token (required for gated models like Llama)"},"placeholder":"hf_..."}]},"addons":{"available":[{"id":"jupyter","name":"JupyterLab","description":"Browser-based Python IDE with GPU monitoring","port":8888,"access_label":"JupyterLab","access_method":"jupyter"},{"id":"code-server","name":"VS Code Server","description":"Full VS Code editor in your browser","port":8443,"access_label":"VS Code","access_method":"code_server"}],"defaults":["jupyter"]}},{"id":"00000000-0000-0000-0000-000000000033","name":"Wan 2.2 Video","slug":"wan22","description":"Text-to-video and image-to-video with Wan 2.2 (beta). 720p/24fps clips from a browser UI — models cached on your volume.","docker_image":"ghcr.io/benpoliquin/templates/wan22@sha256:356a8edfc8a162c1a15d1966a432a4f348898d979b2ccadc9bd50692714f20a4","requires_gpu":true,"min_gpu":"rtx-4090","min_vram_gb":24,"estimated_setup_seconds":300,"category":"video","source_repo":"https://github.com/deepbeepmeep/Wan2GP","billing_model":"per_minute","proxy_port":7860,"monthly_price_cents":null,"volume_gb":100,"base_image_type":"custom","is_active":true,"datasets":[],"gpu_options":[{"id":"9f8e5660-fbff-4ab8-9001-9fbc44c4bc75","template_id":"00000000-0000-0000-0000-000000000033","gpu_type":"rtx-4090","display_name":"RTX 4090","sell_price_per_hr_cents":48,"is_default":true,"sort_order":0,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":300,"available":true,"stock_level":"medium","providers":[{"provider":"runpod","sell_price_per_hr_cents":48,"available":1,"stock_level":"low","region":"","cloud_tier":"COMMUNITY"},{"provider":"vastai","sell_price_per_hr_cents":57,"available":1,"stock_level":"medium","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":59,"available":1,"stock_level":"low","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":61,"available":1,"stock_level":"medium","region":"Texas, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"Pennsylvania, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":70,"available":1,"stock_level":"low","region":"British Columbia, CA","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":85,"available":1,"stock_level":"low","region":"Minnesota, US","cloud_tier":"ON_DEMAND"}],"vram_gb":24,"min_vcpus":12,"max_vcpus":256,"min_ram_gb":31,"max_ram_gb":126},{"id":"e64a9a77-891c-4dae-bcbb-d70aaa610d78","template_id":"00000000-0000-0000-0000-000000000033","gpu_type":"rtx-a6000","display_name":"RTX A6000","sell_price_per_hr_cents":57,"is_default":false,"sort_order":1,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":300,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":57,"available":1,"stock_level":"low","region":"Kansas, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":97,"available":1,"stock_level":"low","region":"","cloud_tier":"SECURE"},{"provider":"lambda_labs","sell_price_per_hr_cents":153,"available":1,"stock_level":"high","region":"us-south-2","cloud_tier":"ON_DEMAND"}],"vram_gb":48,"min_vcpus":14,"max_vcpus":64,"min_ram_gb":63,"max_ram_gb":100},{"id":"8fe2631a-88cc-477e-907a-9f025d3afd8b","template_id":"00000000-0000-0000-0000-000000000033","gpu_type":"l40s","display_name":"L40S","sell_price_per_hr_cents":153,"is_default":false,"sort_order":2,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":300,"available":true,"stock_level":"low","providers":[{"provider":"runpod","sell_price_per_hr_cents":153,"available":1,"stock_level":"low","region":"","cloud_tier":"SECURE"}],"vram_gb":48,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"472bef2b-aa14-4a98-86da-c7b2feb5e5dc","template_id":"00000000-0000-0000-0000-000000000033","gpu_type":"a100-80gb-sxm","display_name":"A100 80GB SXM","sell_price_per_hr_cents":176,"is_default":false,"sort_order":3,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":500,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":176,"available":1,"stock_level":"low","region":"Montana, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":187,"available":1,"stock_level":"low","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":195,"available":1,"stock_level":"low","region":"North Carolina, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":307,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":80,"min_vcpus":128,"max_vcpus":208,"min_ram_gb":126,"max_ram_gb":126}],"frontend_config":{"gpu_guidance":{"l40s":"48 GB VRAM — datacenter-grade; good sustained generation.","rtx-4090":"24 GB VRAM — TI2V-5B with model offloading.","rtx-a6000":"48 GB VRAM — headroom for longer clips and bigger batches.","a100-80gb-sxm":"80 GB VRAM — required for the A14B quality models."},"what_you_get":["Wan2GP web UI — text-to-video and image-to-video","Wan 2.2 TI2V-5B: 720p/24fps 5s clips on 24 GB VRAM","Model weights cached on your volume across restarts"],"env_var_fields":[],"show_gpu_selector":true,"expanded_description":"Open text/image-to-video with Wan 2.2 (Apache 2.0) via the Wan2GP web UI. Pick the TI2V-5B model in the UI for 720p/24fps 5-second clips on a single 24 GB GPU. Model weights (~34 GB) download on first use with in-UI progress and are cached on your persistent volume — restarts skip the download. Beta: expect rough edges; larger A14B models need bigger GPUs.","optional_env_var_fields":[]}},{"id":"00000000-0000-0000-0000-000000000013","name":"vLLM","slug":"vllm","description":"Production-grade LLM serving with OpenAI-compatible API. High throughput, continuous batching, multi-GPU tensor parallelism.","docker_image":"ghcr.io/benpoliquin/templates/vllm@sha256:ee51613906dfad07306fe035f2a19ffb39113891fae5d0b35259b24b6144dca3","requires_gpu":true,"min_gpu":"l40s","min_vram_gb":16,"estimated_setup_seconds":300,"category":"language-models","source_repo":"https://github.com/vllm-project/vllm","billing_model":"per_minute","proxy_port":8080,"monthly_price_cents":null,"volume_gb":50,"base_image_type":"custom","is_active":true,"datasets":[],"gpu_options":[{"id":"e3a410d2-9c85-4e9e-968d-402c439f13af","template_id":"00000000-0000-0000-0000-000000000013","gpu_type":"l40s","display_name":"L40S","sell_price_per_hr_cents":153,"is_default":true,"sort_order":0,"compute_type":"gpu","gpu_count":1,"default_volume_gb":50,"max_volume_gb":200,"available":true,"stock_level":"low","providers":[{"provider":"runpod","sell_price_per_hr_cents":153,"available":1,"stock_level":"low","region":"","cloud_tier":"SECURE"}],"vram_gb":48,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"65a67bd2-d44e-4740-9225-213c472ba8a6","template_id":"00000000-0000-0000-0000-000000000013","gpu_type":"a100-80gb-sxm","display_name":"A100 80GB SXM","sell_price_per_hr_cents":176,"is_default":false,"sort_order":1,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":500,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":176,"available":1,"stock_level":"low","region":"Montana, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":187,"available":1,"stock_level":"low","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":195,"available":1,"stock_level":"low","region":"North Carolina, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":307,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":80,"min_vcpus":128,"max_vcpus":208,"min_ram_gb":126,"max_ram_gb":126},{"id":"2ee46b73-c6d5-4672-be82-98c2d1aace70","template_id":"00000000-0000-0000-0000-000000000013","gpu_type":"h100-sxm","display_name":"H100 SXM","sell_price_per_hr_cents":615,"is_default":false,"sort_order":2,"compute_type":"gpu","gpu_count":1,"default_volume_gb":100,"max_volume_gb":500,"available":true,"stock_level":"high","providers":[{"provider":"runpod","sell_price_per_hr_cents":615,"available":1,"stock_level":"high","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"}],"vram_gb":80,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"d6cf8421-8b45-49fe-85ce-1135d955d270","template_id":"00000000-0000-0000-0000-000000000013","gpu_type":"2xh100-sxm","display_name":"2xH100 SXM","sell_price_per_hr_cents":486,"is_default":false,"sort_order":3,"compute_type":"gpu","gpu_count":2,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":486,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":1230,"available":1,"stock_level":"high","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"}],"vram_gb":160,"min_vcpus":256,"max_vcpus":256,"min_ram_gb":252,"max_ram_gb":252},{"id":"f7c6a5c7-eb26-49b6-935c-2b5459613cd4","template_id":"00000000-0000-0000-0000-000000000013","gpu_type":"2xa100-80gb-sxm","display_name":"2xA100 80GB SXM","sell_price_per_hr_cents":374,"is_default":false,"sort_order":4,"compute_type":"gpu","gpu_count":2,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":374,"available":1,"stock_level":"low","region":"New Jersey, US","cloud_tier":"ON_DEMAND"},{"provider":"vastai","sell_price_per_hr_cents":389,"available":1,"stock_level":"low","region":"North Carolina, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":614,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":160,"min_vcpus":128,"max_vcpus":256,"min_ram_gb":252,"max_ram_gb":252},{"id":"09888b14-1376-442d-afab-f3d49827bcdd","template_id":"00000000-0000-0000-0000-000000000013","gpu_type":"2xl40s","sell_price_per_hr_cents":220,"is_default":false,"sort_order":5,"compute_type":"gpu","gpu_count":2,"default_volume_gb":100,"max_volume_gb":500,"available":false,"stock_level":"none","providers":[],"vram_gb":0,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"6e99cf3d-9f0b-468d-9776-eeb98da4a756","template_id":"00000000-0000-0000-0000-000000000013","gpu_type":"4xa100-80gb-sxm","display_name":"4xA100 80GB SXM","sell_price_per_hr_cents":1227,"is_default":false,"sort_order":6,"compute_type":"gpu","gpu_count":4,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"runpod","sell_price_per_hr_cents":1227,"available":1,"stock_level":"high","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":320,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"e71cad0d-636d-48ff-be5d-8ee1482f1499","template_id":"00000000-0000-0000-0000-000000000013","gpu_type":"4xl40s","sell_price_per_hr_cents":440,"is_default":false,"sort_order":7,"compute_type":"gpu","gpu_count":4,"default_volume_gb":200,"max_volume_gb":1000,"available":false,"stock_level":"none","providers":[],"vram_gb":0,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"4c46910f-db4b-45ca-89b0-5d7dbe673c03","template_id":"00000000-0000-0000-0000-000000000013","gpu_type":"4xh100-sxm","display_name":"4xH100 SXM","sell_price_per_hr_cents":971,"is_default":false,"sort_order":8,"compute_type":"gpu","gpu_count":4,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"high","providers":[{"provider":"vastai","sell_price_per_hr_cents":971,"available":1,"stock_level":"low","region":"Washington, US","cloud_tier":"ON_DEMAND"},{"provider":"runpod","sell_price_per_hr_cents":2459,"available":1,"stock_level":"high","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"}],"vram_gb":320,"min_vcpus":256,"max_vcpus":256,"min_ram_gb":504,"max_ram_gb":504},{"id":"02f96154-a0e9-44ee-bf4f-1cc62c5a09bd","template_id":"00000000-0000-0000-0000-000000000013","gpu_type":"8xa100-80gb-sxm","display_name":"8xA100 80GB SXM","sell_price_per_hr_cents":2453,"is_default":false,"sort_order":9,"compute_type":"gpu","gpu_count":8,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"low","providers":[{"provider":"runpod","sell_price_per_hr_cents":2453,"available":1,"stock_level":"low","region":"United States","datacenter_id":"United States","cloud_tier":"SECURE"}],"vram_gb":640,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0},{"id":"fecc8ae2-5121-4c8b-9f42-516948a90be1","template_id":"00000000-0000-0000-0000-000000000013","gpu_type":"8xh100-sxm","display_name":"8xH100 SXM","sell_price_per_hr_cents":4917,"is_default":false,"sort_order":10,"compute_type":"gpu","gpu_count":8,"default_volume_gb":200,"max_volume_gb":1000,"available":true,"stock_level":"low","providers":[{"provider":"runpod","sell_price_per_hr_cents":4917,"available":1,"stock_level":"low","region":"India, Europe","datacenter_id":"India","cloud_tier":"SECURE"}],"vram_gb":640,"min_vcpus":0,"max_vcpus":0,"min_ram_gb":0,"max_ram_gb":0}],"frontend_config":{"gpu_badges":{"a100-80gb-sxm":"Best throughput per dollar"},"gpu_guidance":{"l40s":"48 GB VRAM — best cost-per-token. Runs Qwen3-8B, Phi-4 Mini, Mistral Small 3.2, gpt-oss 20B, Magistral 24B.","2xl40s":"96 GB VRAM — cost-effective multi-GPU inference. Run multiple mid-tier models in parallel.","4xl40s":"192 GB VRAM — budget multi-GPU. Mid-tier 70B quantized or parallel serving.","h100-sxm":"80 GB VRAM — 2-3x faster than A100. Lowest latency for Qwen3-32B, gpt-oss 120B production serving.","2xh100-sxm":"160 GB VRAM — Llama 3.3 70B FP16, or DeepSeek R1 Distill 32B with long context.","4xh100-sxm":"320 GB VRAM — Llama 4 Scout (109B MoE) or Qwen3-235B FP8 with tensor parallelism.","8xh100-sxm":"640 GB VRAM — DGX-class inference. DeepSeek V3.1 671B, GLM-4.6, Llama 4 Maverick at full precision.","a100-80gb-sxm":"80 GB VRAM — single-GPU sweet spot for 24B-32B dense (Qwen3-32B, Gemma 3 27B, Qwen3-Coder 30B) or gpt-oss 120B MoE (MXFP4).","2xa100-80gb-sxm":"160 GB VRAM — serve Llama 3.3 70B FP16 or 30B+ models with concurrent requests.","4xa100-80gb-sxm":"320 GB VRAM — serve Qwen3-235B-A22B (FP8) or Llama 4 Scout with long context.","8xa100-80gb-sxm":"640 GB VRAM — frontier MoE (DeepSeek V3.1, Llama 4 Maverick) with FP8 quantization."},"what_you_get":["OpenAI-compatible API (drop-in replacement)","ChatGPT-like web UI (Open WebUI)","Continuous batching for high throughput","Multi-GPU tensor parallelism (auto-detected)","JupyterLab for scripting and API testing"],"env_var_fields":[{"key":"VLLM_MODEL","label":"Model","options":[{"label":"Qwen 3 8B — ~16 GB VRAM (thinking mode)","value":"Qwen/Qwen3-8B"},{"label":"Qwen3.8 27B — vision, coding, and reasoning (80 GB GPU recommended)","value":"Qwen/Qwen3.8-27B","recommended_gpu":"a100-80gb-sxm"},{"label":"Gemma 4 26B-A4B — multimodal MoE, 4B active (80 GB GPU recommended)","value":"google/gemma-4-26B-A4B-it","recommended_gpu":"a100-80gb-sxm"},{"label":"Phi-4 Mini 3.8B — ~8 GB VRAM","value":"microsoft/Phi-4-mini-instruct"},{"label":"Llama 3.1 8B Instruct — ~18 GB VRAM (needs HF token)","value":"meta-llama/Llama-3.1-8B-Instruct"},{"label":"Gemma 3 12B IT — ~24 GB VRAM (vision, needs HF token)","value":"google/gemma-3-12b-it"},{"label":"Qwen3-Coder 30B MoE — 3.3B active, 256K ctx, ~60 GB VRAM","value":"Qwen/Qwen3-Coder-30B-A3B-Instruct","recommended_gpu":"a100-80gb-sxm"},{"label":"Qwen 2.5 Coder 32B — ~64 GB VRAM","value":"Qwen/Qwen2.5-Coder-32B-Instruct","recommended_gpu":"a100-80gb-sxm"},{"label":"gpt-oss 20B MoE — fits 16 GB (MXFP4), reasoning/agents","value":"openai/gpt-oss-20b"},{"label":"DeepSeek R1 Distill 32B — ~64 GB VRAM (reasoning)","value":"deepseek-ai/DeepSeek-R1-Distill-Qwen-32B","recommended_gpu":"a100-80gb-sxm"},{"label":"Magistral Small 24B — ~48 GB VRAM (reasoning)","value":"mistralai/Magistral-Small-2509"},{"label":"Qwen 3 32B — ~64 GB VRAM","value":"Qwen/Qwen3-32B","recommended_gpu":"a100-80gb-sxm"},{"label":"Gemma 3 27B IT — ~54 GB VRAM (vision, needs HF token)","value":"google/gemma-3-27b-it","recommended_gpu":"a100-80gb-sxm"},{"label":"Mistral Small 3.2 24B — ~48 GB VRAM (vision)","value":"mistralai/Mistral-Small-3.2-24B-Instruct-2506"},{"label":"Llama 3.3 70B Instruct — ~140 GB VRAM (needs HF token)","value":"meta-llama/Llama-3.3-70B-Instruct","recommended_gpu":"2xh100-sxm"},{"label":"gpt-oss 120B MoE — fits single 80 GB (MXFP4), reasoning","value":"openai/gpt-oss-120b","recommended_gpu":"a100-80gb-sxm"},{"label":"Llama 4 Scout 17B×16E MoE — 109B total, 10M ctx (needs HF token)","value":"meta-llama/Llama-4-Scout-17B-16E-Instruct","recommended_gpu":"4xh100-sxm"},{"label":"Qwen 3 235B-A22B MoE — 22B active, 256K ctx, frontier-class","value":"Qwen/Qwen3-235B-A22B-Instruct-2507","recommended_gpu":"4xh100-sxm"},{"label":"GLM-4.6 357B — 200K ctx, MIT license","value":"zai-org/GLM-4.6","recommended_gpu":"8xh100-sxm"},{"label":"DeepSeek V3.1 671B MoE — 37B active, frontier-class","value":"deepseek-ai/DeepSeek-V3.1","recommended_gpu":"8xh100-sxm"},{"label":"Llama 4 Maverick 17B×128E MoE — 400B total (needs HF token)","value":"meta-llama/Llama-4-Maverick-17B-128E-Instruct","recommended_gpu":"8xh100-sxm"}],"required":true,"placeholder":"Qwen/Qwen3-8B"}],"contact_for_more":"Need custom configurations or dedicated clusters?","show_gpu_selector":true,"expanded_description":"Production-grade LLM inference with vLLM. Serves any HuggingFace model via an OpenAI-compatible API — swap your base_url and your existing code works. Includes Open WebUI for interactive chat and JupyterLab for scripting. Features continuous batching, PagedAttention, and multi-GPU tensor parallelism for high-throughput serving.","optional_env_var_fields":[{"key":"HF_TOKEN","label":"HuggingFace Token","required":false,"help_link":{"url":"https://huggingface.co/settings/tokens","text":"Get a token (required for Llama, Gemma)"},"placeholder":"hf_..."},{"key":"VLLM_ARGS","label":"Extra vLLM Arguments","required":false,"help_link":{"url":"https://docs.vllm.ai/en/latest/serving/openai_compatible_server.html","text":"All vLLM options"},"sensitive":false,"placeholder":"--max-model-len 8192"}]},"addons":{"available":[{"id":"jupyter","name":"JupyterLab","description":"Browser-based Python IDE with GPU monitoring","port":8888,"access_label":"JupyterLab","access_method":"jupyter"},{"id":"code-server","name":"VS Code Server","description":"Full VS Code editor in your browser","port":8443,"access_label":"VS Code","access_method":"code_server"}],"defaults":["jupyter"]}}]}
