Models.dev
Anomaly Co (OpenCode team) · Other
Models.dev is a free, open-source database for comparing AI models, their providers, specs, and pricing in one place. It gathers model facts that normally sit scattered across a dozen vendor pages and serves them as a browsable table and a public model pricing API. Developers use it for AI model comparison before picking a provider, and tool builders pull the same data into cost calculators or model routers. Want the short version? It's an open-source model catalog you can read, fork, or query for free.
About Models.dev
What Is Models.dev
Models.dev is an open-source database of AI model specifications, pricing, and capabilities, run by the team behind the OpenCode coding agent and hosted under the Anomaly Co organization on GitHub. The project exists because, as its own README puts it, there's no single database covering every available AI model. So the team built one and opened it to contributions.
The database answers the questions people ask most about a model: what it costs per million tokens, how large its context window is, which providers serve it, and whether it supports reasoning, tool calls, or structured output. It holds AI model specs that would otherwise take an afternoon of tab-hopping to assemble. All of it lives in the repository as TOML files, organized by provider and model, and feeds both the website and the API.
The main limit is freshness. Because the catalog is community-maintained rather than vendor-published, a provider can quietly change prices or retire a model long before anyone gets around to updating the entry. Treat Models.dev as a fast starting point, then confirm anything critical at the source.
Getting Started
- Open models.dev in any browser. No account, login, or subscription is needed.
- Browse or filter the model table to compare context size, pricing, and supported features side by side.
- Click any model to see its model ID, the providers that serve it, and its capability flags.
- For programmatic use, pull the catalog with
curl https://models.dev/api.jsonand read the JSON in your own code. - If you spot a wrong price or a missing model, submit a pull request against the GitHub repository.
Product Information
A quick look at Models.dev's pricing, supported platforms, and performance.
Best for
The users, tasks, and scenarios where this tool fits best.
Users
- Developers
- Tool builders
- Technical writers and researchers
Tasks
- Model comparison
- Pricing lookups
- Data integration
Scenarios
- Choosing between a cheap and a premium model for a summarization feature.
- Due diligence before adopting a new provider's API.
- Keeping an internal model list in sync with the wider market.
Key features
One Table for Every Model
Models.dev lists models in a single filterable table instead of spreading them across vendor sites, which means you can scan names, providers, context sizes, and pricing without leaving the page. For anyone sizing up options across several labs, that alone removes most of the busywork that used to come with a comparison.
A Public JSON API
The data serves machines as well as people. Hitting https://models.dev/api.json returns the whole catalog as structured JSON, complete with pricing and capability flags, and there are separate feeds like models.json and catalog.json for narrower needs. If you maintain a router or a cost tool, you can fetch on a schedule and skip scraping.
Provider-Agnostic Model Metadata
The models.json feed holds facts about a model itself, separate from whichever provider serves it, such as release date, knowledge cutoff, and supported modalities. That separation matters when the same underlying model shows up across several providers that each apply their own context limits and pricing.
Provider Logos and IDs
Each provider has a logo served as an SVG at a predictable URL, and models use the ID format from the AI SDK. Drop-in identifiers mean the data lines up with what many JavaScript and TypeScript projects already expect.
Open Source Under MIT
The whole dataset is public, licensed under MIT, and stored as readable TOML files, so you can fork it, trim it down to the providers you actually care about, or serve your own internal copy behind a firewall. No vendor holds the keys.
Community Contributions
Because the data lives in a repo, fixes follow the pull-request flow developers already know. Spot a stale price or a missing model, open a PR, and the database improves as the market moves.
Pros and cons
Pros
- Free to browse and query, with no account or paywall in the way.
- Structured data and a public API make it easy to reuse in other tools.
- Open-source under MIT, so you can inspect, fork, or self-host the catalog.
- Tracks pricing, context limits, and capability flags in one comparable table.
- Provider-agnostic metadata separates model facts from serving details.
Cons
- Community-maintained entries can lag behind a fast release schedule, so prices may be stale.
- No benchmark charts or sample outputs, so it won't tell you which model actually performs better.
- Coverage is broad but not exhaustive, and obscure providers may be missing.
- Accuracy depends on contributors, so critical numbers need a check at the source.
Frequently asked questions
It's a reference for comparing AI models, their providers, specs, and pricing in one place. Developers use it before picking a provider or wiring an API, and tool builders pull the data into their own software.
Related content
Explore related tools, skills, and articles for Models.dev.
Models.dev Alternatives
AI 3D Model Generator
AI 3D Model · OtherAI 3D Model Generator is a browser-based tool that turns text prompts, single images, or multi-view photos into finished 3D models without any modeling software. It's built for 3D artists, game developers, and creators who need usable assets fast, and it exports to GLB, OBJ, STL, FBX, and other standard formats so the results drop straight into Blender, Unity, or a slicer. Generation runs on credits. The free tier is enough to test the workflow before you pay.
Airmusic
AirMusic · OtherAirMusic is a browser-based AI song maker that turns a text prompt or your own lyrics into a finished song, complete with vocals and instruments. You pick a style, add optional lyrics, choose a model, then preview and download the result as MP3 or WAV. It's built for people who want usable audio fast rather than musicians working inside a DAW.
Nyckel
Nyckel · Coding · OtherNyckel is a managed machine learning platform that turns labeled examples into production-ready prediction APIs. You upload a handful of samples, and Nyckel trains, hosts, and serves a custom model behind a stable REST endpoint for classification, object detection, or similarity search. Think of it as an AI classification API that doubles as an auto ML platform: the service handles the model work, you handle the labels. It's built for teams that need consistent AI decisions at scale without hiring ML specialists or managing GPUs.