# GPU Ninja

> GPU Ninja is an open database of GPUs: chips, boards, chassis, AI models, benchmarks and platforms, with prices, stock and compatibility. Every page has a plaintext `.txt` companion, and the data is a free JSON API at https://gpuninja.com/api/v1 with no key and permissive CORS.

Compatibility answers ("does this board fit this case", "can this board run this model") are precomputed and published under each entity's API path. An MCP server for agents is documented at https://gpuninja.com/docs/mcp.txt.

## Site
- [GPU Ninja home](https://gpuninja.com/index.txt): What GPU Ninja covers and where to start, in plain text

## Docs
- [Documentation](https://gpuninja.com/docs.txt): What GPU Ninja is, how to use it, and a map of every part of it, from the website and JSON API to the MCP server, desktop app and daily digest.
- [Accessibility](https://gpuninja.com/docs/accessibility.txt): GPU Ninja's accessibility statement - our WCAG 2.2 AA target, what we do to meet it, the known limits, and how to report a problem.
- [Your account](https://gpuninja.com/docs/account.txt): Signing in to app.gpuninja.com, the digest subscription, unsubscribing, connected apps, downloading your data and deleting your account.
- [Public API](https://gpuninja.com/docs/api.txt): Read GPU Ninja's catalogue as static JSON. No key, no rate limit, cached for five minutes, with plain-text companions for every page.
- [Caching, redirects and errors](https://gpuninja.com/docs/api/caching-and-errors.txt): Response headers, ETag revalidation, CORS limits, redirects for renames and trailing slashes, and what a 404 or 405 looks like.
- [Changes feed](https://gpuninja.com/docs/api/changes.txt): The machine-readable record of every data change in the last 30 days, the monthly archives, and how to poll them to keep a local copy current.
- [Collections and entities](https://gpuninja.com/docs/api/collections.txt): The nine collections, how paging works, and the fields of every entity summary and detail document, with provenance and frozen records.
- [Compatibility](https://gpuninja.com/docs/api/compatibility.txt): Which cases a card fits, which cards fit a case, and which AI models a card can run, with the exact rules behind each answer.
- [API reference](https://gpuninja.com/docs/api/reference.txt): Every endpoint, parameter and response of the public JSON API
- [Search index](https://gpuninja.com/docs/api/search.txt): The compact search index of chips, boards, cases, models, motherboards and risers, its fields per kind, and how to search it with the open-source library.
- [GPU benchmark](https://gpuninja.com/docs/browser-benchmark.txt): How the browser benchmark tests your GPU, how it decides good to go, tight or won't run for a workload, and what it keeps.
- [Buyer's guide and site pages](https://gpuninja.com/docs/buyers-guide.txt): The catalog, detail and compatibility pages, the buyer's guide's filters, URL parameters, comparison and saved plan, and the plain-text companion of every page.
- [Using GPU Ninja in ChatGPT](https://gpuninja.com/docs/chatgpt.txt): Enable GPU Ninja's sign-in-free connector in ChatGPT, and how it differs from the consumer MCP server that Claude uses.
- [Using GPU Ninja in Claude](https://gpuninja.com/docs/claude-connector.txt): Add GPU Ninja to Claude from the Connectors Directory, as a custom connector, or in Claude Code, and sign in with a free account.
- [Data provenance](https://gpuninja.com/docs/data-provenance.txt): How GPU Ninja's data is researched, verified and reviewed, and how site owners can identify or block the research bot, GPUNinjaResearchBot.
- [Data](https://gpuninja.com/docs/data.txt): What GPU Ninja's data covers, what discontinued and frozen mean, how freshness and provenance work, how prices are stored, and how to report an error.
- [Desktop app changelog](https://gpuninja.com/docs/desktop-changelog.txt): Every release of the GPU Ninja desktop app for macOS, Windows and Linux, newest first, with what changed in each.
- [Desktop app](https://gpuninja.com/docs/desktop.txt): Download GPU Ninja for macOS, Windows or Linux. It detects your GPU locally, works offline and alerts you when cards you watch change.
- [Daily digest](https://gpuninja.com/docs/digest.txt): What the daily digest contains, when it is sent, where to read it on the web and by RSS, and how to subscribe and unsubscribe.
- [Feeds and plaintext](https://gpuninja.com/docs/feeds.txt): Every feed and machine-readable file GPU Ninja publishes, from blog and digest feeds to llms.txt, the sitemap and the .txt page companions.
- [Getting started](https://gpuninja.com/docs/getting-started.txt): The fastest way to use GPU Ninja, whether you want JSON, a search, an AI assistant, the desktop app or the daily digest.
- [Using the MCP servers](https://gpuninja.com/docs/mcp.txt): Connect an AI assistant to GPU Ninja over MCP, sign in to the consumer server or use the sign-in-free connector, and see what each tool does.
- [Open source](https://gpuninja.com/docs/open-source.txt): The Apache 2.0 gpuninja-mcp repository, what it contains, how to run your own copy of the MCP server, and how to contribute.
- [Support](https://gpuninja.com/docs/support.txt): How to get help with GPU Ninja, report a bug or a data error, and answers to the most common problems with sign-in, the digest and connectors.

## API
- [Compact search index](https://gpuninja.com/api/v1/index): Every chip, board, chassis, model, motherboard and riser with its slug, for search and lookups
- [Changes](https://gpuninja.com/api/v1/changes): Data changes merged in the last 30 days, newest first
- [OpenAPI document](https://gpuninja.com/api/v1/openapi.json): The OpenAPI description of every endpoint of the public JSON API
- [Chips](https://gpuninja.com/api/v1/chips): Paginated, 50 per page; discontinued chips are left out, and every one is at https://gpuninja.com/api/v1/chips/all; each item's detail document is at https://gpuninja.com/api/v1/chips/{slug}
- [Boards](https://gpuninja.com/api/v1/boards): Paginated, 50 per page; discontinued boards are left out, and every one is at https://gpuninja.com/api/v1/boards/all; each item's detail document is at https://gpuninja.com/api/v1/boards/{slug}
- [Chassis](https://gpuninja.com/api/v1/chassis): Paginated, 50 per page; discontinued chassis are left out, and every one is at https://gpuninja.com/api/v1/chassis/all; each item's detail document is at https://gpuninja.com/api/v1/chassis/{slug}
- [Models](https://gpuninja.com/api/v1/models): Paginated, 50 per page; each item's detail document is at https://gpuninja.com/api/v1/models/{slug}
- [Platforms](https://gpuninja.com/api/v1/platforms): Paginated, 50 per page; each item's detail document is at https://gpuninja.com/api/v1/platforms/{slug}
- [Benchmarks](https://gpuninja.com/api/v1/benchmarks): Paginated, 50 per page; each item's detail document is at https://gpuninja.com/api/v1/benchmarks/{slug}
- [Motherboards](https://gpuninja.com/api/v1/motherboards): Paginated, 50 per page; discontinued motherboards are left out, and every one is at https://gpuninja.com/api/v1/motherboards/all; each item's detail document is at https://gpuninja.com/api/v1/motherboards/{slug}
- [Risers](https://gpuninja.com/api/v1/risers): Paginated, 50 per page; discontinued risers are left out, and every one is at https://gpuninja.com/api/v1/risers/all; each item's detail document is at https://gpuninja.com/api/v1/risers/{slug}
- [Sockets](https://gpuninja.com/api/v1/sockets): Paginated, 50 per page; each item's detail document is at https://gpuninja.com/api/v1/sockets/{slug}

## Chips
- [All chips](https://gpuninja.com/chips.txt): Catalog of 1,185 chips; each chip's page is https://gpuninja.com/chips/{slug}.txt

## Boards
- [All boards](https://gpuninja.com/boards.txt): Catalog of 38 boards, 50 per page; discontinued boards are left out
- [Radeon RX 9070 XT Taichi 16GB OC](https://gpuninja.com/boards/asrock-taichi-9070-xt-oc.txt)
- [PRIME GeForce RTX 5070 OC Edition](https://gpuninja.com/boards/asus-prime-5070-oc.txt)
- [Prime Radeon RX 9070 XT OC Edition](https://gpuninja.com/boards/asus-prime-9070-xt-oc.txt)
- [ProArt GeForce RTX 4080 SUPER](https://gpuninja.com/boards/asus-proart-4080-super.txt)
- [ROG Astral GeForce RTX 5090 OC Edition](https://gpuninja.com/boards/asus-rog-astral-5090-oc.txt)
- [ROG Strix GeForce RTX 4090 OC Edition](https://gpuninja.com/boards/asus-rog-strix-4090-oc.txt)
- [ROG Strix GeForce RTX 4090](https://gpuninja.com/boards/asus-rog-strix-4090.txt)
- [TUF Gaming GeForce RTX 4090 OC Edition](https://gpuninja.com/boards/asus-tuf-gaming-4090-oc.txt)
- [TUF Gaming GeForce RTX 5070 Ti OC Edition](https://gpuninja.com/boards/asus-tuf-gaming-5070-ti-oc.txt)
- [TUF Gaming GeForce RTX 5080 OC Edition](https://gpuninja.com/boards/asus-tuf-gaming-5080-oc.txt)
- [TUF Gaming Radeon RX 9070 XT OC Edition](https://gpuninja.com/boards/asus-tuf-gaming-9070-xt-oc.txt)
- [GeForce RTX 4090 GAMING OC 24G](https://gpuninja.com/boards/gigabyte-gaming-4090-oc.txt)
- [Radeon RX 9070 XT GAMING OC 16G](https://gpuninja.com/boards/gigabyte-gaming-9070-xt-oc.txt)
- [Arc B580 Limited Edition](https://gpuninja.com/boards/intel-b580-limited-edition.txt)
- [GeForce RTX 5080 16G GAMING TRIO OC](https://gpuninja.com/boards/msi-gaming-trio-5080-oc.txt)
- [GeForce RTX 5090 32G GAMING TRIO OC](https://gpuninja.com/boards/msi-gaming-trio-5090-oc.txt)
- [GeForce RTX 4090 GAMING X TRIO 24G](https://gpuninja.com/boards/msi-gaming-x-trio-4090.txt)
- [GeForce RTX 4090 SUPRIM X 24G](https://gpuninja.com/boards/msi-suprim-x-4090.txt)
- [GeForce RTX 4070 SUPER 12G VENTUS 2X OC](https://gpuninja.com/boards/msi-ventus-2x-4070-super-oc.txt)
- [GeForce RTX 5060 Ti 16G VENTUS 2X OC PLUS](https://gpuninja.com/boards/msi-ventus-2x-5060-ti-16-gb-oc-plus.txt)
- [GeForce RTX 5070 12G VENTUS 2X OC](https://gpuninja.com/boards/msi-ventus-2x-5070-oc.txt)
- [GeForce RTX 5070 Ti 16G VENTUS 3X OC](https://gpuninja.com/boards/msi-ventus-3x-5070-ti-oc.txt)
- [GeForce RTX 5070 Founders Edition](https://gpuninja.com/boards/nvidia-5070-founders-edition.txt)
- [GeForce RTX 5080 Founders Edition](https://gpuninja.com/boards/nvidia-5080-founders-edition.txt)
- [GeForce RTX 5090 Founders Edition](https://gpuninja.com/boards/nvidia-5090-founders-edition.txt)
- [Hellhound AMD Radeon RX 9070 XT 16GB GDDR6](https://gpuninja.com/boards/powercolor-hellhound-9070-xt.txt)
- [Reaper AMD Radeon RX 9070 XT 16GB GDDR6](https://gpuninja.com/boards/powercolor-reaper-9070-xt.txt)
- [Red Devil AMD Radeon RX 9070 XT 16GB GDDR6](https://gpuninja.com/boards/powercolor-red-devil-9070-xt.txt)
- [NITRO+ AMD Radeon RX 7900 XTX Vapor-X 24GB](https://gpuninja.com/boards/sapphire-nitro-7900-xtx-vapor-x.txt)
- [NITRO+ AMD Radeon RX 9070 XT OC](https://gpuninja.com/boards/sapphire-nitro-9070-xt-oc.txt)
- [PULSE AMD Radeon RX 7800 XT 16GB](https://gpuninja.com/boards/sapphire-pulse-7800-xt.txt)
- [PULSE AMD Radeon RX 7900 XTX 24GB](https://gpuninja.com/boards/sapphire-pulse-7900-xtx.txt)
- [PULSE AMD Radeon RX 9060 XT OC 16GB](https://gpuninja.com/boards/sapphire-pulse-9060-xt-16-gb-oc.txt)
- [PULSE AMD Radeon RX 9070 XT](https://gpuninja.com/boards/sapphire-pulse-9070-xt.txt)
- [PULSE AMD Radeon RX 9070](https://gpuninja.com/boards/sapphire-pulse-9070.txt)
- [PURE AMD Radeon RX 9070 XT OC](https://gpuninja.com/boards/sapphire-pure-9070-xt-oc.txt)
- [Mercury AMD Radeon RX 9070XT OC Magnetic Air Edition with RGB](https://gpuninja.com/boards/xfx-mercury-9070-xt-oc-magnetic-air.txt)
- [Swift AMD Radeon RX 9070XT Triple Fan Gaming Edition](https://gpuninja.com/boards/xfx-swift-9070-xt.txt)

## Chassis
- [All chassis](https://gpuninja.com/chassis.txt): Catalog of 27 chassis, 50 per page; discontinued chassis are left out
- [Dark Base 701](https://gpuninja.com/chassis/be-quiet-dark-base-701.txt)
- [Pure Base 500DX](https://gpuninja.com/chassis/be-quiet-pure-base-500dx.txt): Discontinued
- [Shadow Base 800 DX](https://gpuninja.com/chassis/be-quiet-shadow-base-800-dx.txt)
- [HAF 700 EVO](https://gpuninja.com/chassis/cooler-master-haf-700-evo.txt)
- [MasterFrame 700](https://gpuninja.com/chassis/cooler-master-masterframe-700.txt)
- [2500X](https://gpuninja.com/chassis/corsair-2500x.txt)
- [4000D Airflow](https://gpuninja.com/chassis/corsair-4000d-airflow.txt)
- [5000D Airflow](https://gpuninja.com/chassis/corsair-5000d-airflow.txt)
- [7000D Airflow](https://gpuninja.com/chassis/corsair-7000d-airflow.txt)
- [Meshify 2](https://gpuninja.com/chassis/fractal-design-meshify-2.txt)
- [North XL](https://gpuninja.com/chassis/fractal-design-north-xl.txt)
- [North](https://gpuninja.com/chassis/fractal-design-north.txt)
- [Pop Air](https://gpuninja.com/chassis/fractal-design-pop-air.txt)
- [Torrent](https://gpuninja.com/chassis/fractal-design-torrent.txt)
- [D31 MESH](https://gpuninja.com/chassis/jonsbo-d31-mesh.txt)
- [A3-mATX](https://gpuninja.com/chassis/lian-li-a3-matx.txt)
- [Lancool 216](https://gpuninja.com/chassis/lian-li-lancool-216.txt)
- [Lancool III](https://gpuninja.com/chassis/lian-li-lancool-iii.txt)
- [O11 Dynamic EVO XL](https://gpuninja.com/chassis/lian-li-o11-dynamic-evo-xl.txt)
- [O11 Dynamic EVO](https://gpuninja.com/chassis/lian-li-o11-dynamic-evo.txt): Discontinued
- [O11 Vision](https://gpuninja.com/chassis/lian-li-o11-vision.txt)
- [H5 Flow (2024)](https://gpuninja.com/chassis/nzxt-h5-flow-2024.txt)
- [H6 Flow](https://gpuninja.com/chassis/nzxt-h6-flow.txt)
- [H7 Flow (2024)](https://gpuninja.com/chassis/nzxt-h7-flow-2024.txt)
- [H9 Flow (2025)](https://gpuninja.com/chassis/nzxt-h9-flow-2025.txt)
- [Eclipse G360A](https://gpuninja.com/chassis/phanteks-eclipse-g360a.txt): Discontinued
- [NV5](https://gpuninja.com/chassis/phanteks-nv5.txt)
- [NV7](https://gpuninja.com/chassis/phanteks-nv7.txt)
- [Meshroom D](https://gpuninja.com/chassis/ssupd-meshroom-d.txt)
- [Core P3 TG Pro](https://gpuninja.com/chassis/thermaltake-core-p3-tg-pro.txt)

## Models
- [All models](https://gpuninja.com/models.txt): Catalog of 32 models, 50 per page
- [DeepSeek-R1-Distill-Llama-70B (Q4_K_M)](https://gpuninja.com/models/deepseek-r1-distill-llama-70b-q4-k-m.txt)
- [DeepSeek-R1-Distill-Llama-8B (Q4_K_M)](https://gpuninja.com/models/deepseek-r1-distill-llama-8b-q4-k-m.txt)
- [DeepSeek-R1-Distill-Qwen-14B (Q4_K_M)](https://gpuninja.com/models/deepseek-r1-distill-qwen-14b-q4-k-m.txt)
- [DeepSeek-R1-Distill-Qwen-32B (Q4_K_M)](https://gpuninja.com/models/deepseek-r1-distill-qwen-32b-q4-k-m.txt)
- [DeepSeek-R1-Distill-Qwen-7B (Q4_K_M)](https://gpuninja.com/models/deepseek-r1-distill-qwen-7b-q4-k-m.txt)
- [Gemma 3 12B IT (Q4_K_M)](https://gpuninja.com/models/gemma-3-12b-it-q4-k-m.txt)
- [Gemma 3 27B IT (Q4_K_M)](https://gpuninja.com/models/gemma-3-27b-it-q4-k-m.txt)
- [Gemma 3 4B IT (Q4_K_M)](https://gpuninja.com/models/gemma-3-4b-it-q4-k-m.txt)
- [gpt-oss-120b (MXFP4)](https://gpuninja.com/models/gpt-oss-120b-mxfp4.txt)
- [gpt-oss-20b (MXFP4)](https://gpuninja.com/models/gpt-oss-20b-mxfp4.txt)
- [Llama 3.1 70B Instruct (Q4_K_M)](https://gpuninja.com/models/llama-3-1-70b-instruct-q4-k-m.txt)
- [Llama 3.1 8B Instruct (Q4_K_M)](https://gpuninja.com/models/llama-3-1-8b-instruct-q4-k-m.txt)
- [Llama 3.2 1B Instruct (Q4_K_M)](https://gpuninja.com/models/llama-3-2-1b-instruct-q4-k-m.txt)
- [Llama 3.2 3B Instruct (Q4_K_M)](https://gpuninja.com/models/llama-3-2-3b-instruct-q4-k-m.txt)
- [Llama 3.3 70B Instruct (Q4_K_M)](https://gpuninja.com/models/llama-3-3-70b-instruct-q4-k-m.txt)
- [Mistral 7B Instruct v0.3 (Q4_K_M)](https://gpuninja.com/models/mistral-7b-instruct-v0-3-q4-k-m.txt)
- [Mistral Nemo Instruct 2407 (Q4_K_M)](https://gpuninja.com/models/mistral-nemo-instruct-2407-q4-k-m.txt)
- [Mistral Small 3.1 24B Instruct 2503 (Q4_K_M)](https://gpuninja.com/models/mistral-small-3-1-24b-instruct-2503-q4-k-m.txt)
- [Mixtral 8x7B Instruct v0.1 (Q4_K_M)](https://gpuninja.com/models/mixtral-8x7b-instruct-v0-1-q4-k-m.txt)
- [Phi-4-mini-instruct (Q4_K_M)](https://gpuninja.com/models/phi-4-mini-instruct-q4-k-m.txt)
- [Phi-4 (Q4_K_M)](https://gpuninja.com/models/phi-4-q4-k-m.txt)
- [Qwen2.5 14B Instruct (Q4_K_M)](https://gpuninja.com/models/qwen2-5-14b-instruct-q4-k-m.txt)
- [Qwen2.5 32B Instruct (Q4_K_M)](https://gpuninja.com/models/qwen2-5-32b-instruct-q4-k-m.txt)
- [Qwen2.5 72B Instruct (Q4_K_M)](https://gpuninja.com/models/qwen2-5-72b-instruct-q4-k-m.txt)
- [Qwen2.5 7B Instruct (Q4_K_M)](https://gpuninja.com/models/qwen2-5-7b-instruct-q4-k-m.txt)
- [Qwen2.5-Coder 32B Instruct (Q4_K_M)](https://gpuninja.com/models/qwen2-5-coder-32b-instruct-q4-k-m.txt)
- [Qwen3 14B (Q4_K_M)](https://gpuninja.com/models/qwen3-14b-q4-k-m.txt)
- [Qwen3 235B-A22B (Q4_K_M)](https://gpuninja.com/models/qwen3-235b-a22b-q4-k-m.txt)
- [Qwen3 30B-A3B (Q4_K_M)](https://gpuninja.com/models/qwen3-30b-a3b-q4-k-m.txt)
- [Qwen3 32B (Q4_K_M)](https://gpuninja.com/models/qwen3-32b-q4-k-m.txt)
- [Qwen3 4B (Q4_K_M)](https://gpuninja.com/models/qwen3-4b-q4-k-m.txt)
- [Qwen3 8B (Q4_K_M)](https://gpuninja.com/models/qwen3-8b-q4-k-m.txt)

## Blog
- [Blog feed (RSS)](https://gpuninja.com/blog/feed.xml): Every published post, newest first
- [Coming soon](https://gpuninja.com/blog/coming-soon.txt): GPU Ninja is a GPU catalogue and a set of tools for choosing hardware to run AI models. The catalogue goes live soon.

## Optional
- [Legal](https://gpuninja.com/legal.txt): GPU Ninja's legal pages, including the licence for the catalogue data and the API.
- [Data licence](https://gpuninja.com/legal/data-licence.txt): GPU Ninja's catalogue data is CC BY 4.0. Use it for anything, including commercial apps, as long as you credit GPU Ninja.
- [Privacy](https://gpuninja.com/legal/privacy.txt): Privacy information for GPU Ninja’s website and early browser tools.