Open Registry
for AI Infrastructure

Claude Opus 5.5

Anthropic's Opus-class model for long-running agentic coding and knowledge work, succeeding Claude Opus 5 at a lower price. Particularly strong at multi-step changes in large codebases and sustained autonomous tasks, with always-on adaptive thinking steered by an effort parameter, a 1M-token context window, and up to 128K output tokens. Anthropic's recommended starting point for most workloads.

GPT-6 Astra

OpenAI's most capable frontier model for the hardest end-to-end work across complex reasoning, coding, computer use, browsing, research, scientific analysis, cybersecurity defense, and professional document creation. Supports long-context multimodal input, structured outputs, function calling, reasoning effort controls, MCP, skills, computer use, hosted shell, code interpreter, web search, file search, and apply-patch workflows through the Responses API.

GPT-6.1 Sol

OpenAI's upgrade to GPT-6 Sol, delivering near-Astra performance for complex work at a lower cost. Suited for agentic coding, computer use, and document-heavy professional work, with multimodal input, function calling, reasoning effort controls, a 1.05M-token context window, up to 128K output tokens, and cheaper cached input than GPT-6 Sol.

Gemini 4 Argon

Google's flagship reasoning model for software engineering, mathematics, science, and autonomous agents, generally available in Gemini API and Google AI Studio since October 13, 2026. Input, thinking, and output share a 1,048,576-token window, with long responses supported through background execution. Standard API pricing is $2 input and $10 output per million tokens through November 30, 2026; Google has announced $4 input and $20 output from December 1. No free API tier is available.

Kimi K3.1 Pro

Moonshot AI's K3.1 Pro multimodal reasoning model, post-trained from Kimi K3 for coding, 3D creation, financial research and tool-using agents. The documented API supports text, image and video input, structured output and a 1,048,576-token context. Downloadable weights are announced for November 4, 2026; the exact weight license is not yet published. A separately priced Highspeed endpoint is also documented.

Qwen3 Coder 30B A3B Instruct

Alibaba's Qwen3-Coder-30B-A3B-Instruct, a mixture-of-experts coding model with 30.5B total and 3.3B activated parameters. Built for agentic coding and tool use, with native 262,144-token context extendable to 1M tokens with YaRN.

Mistral Large 4

Mistral AI's multimodal model for reasoning, coding, and agentic workloads, with 1.05T total and 49B active parameters. Public preview with a 1M-token context; downloadable weights and the exact license are still marked as coming soon.

GLM-5.3-Prime

The high-speed variant of Z.ai's GLM-5.3, inheriting its full capabilities while delivering 1.5-2x the output throughput through inference acceleration. Text input and output with reasoning, tool use, and a 1M-token context window.

GPT-6.1 Terra

OpenAI's efficient GPT-6.1 model for image-based reasoning, screen interaction, writing and focused coding tasks. Accepts text and images and generates text, with a 1,050,000-token context and up to 128,000 output tokens. Reasoning effort ranges from none to max, with medium as the default. Streaming and structured outputs are supported; tool calling requires the Responses API, not Chat Completions. Fine-tuning, audio and video are not supported. Standard prompts above 272K input tokens cost 2x the input/cache rates and 1.5x the output rate for the entire request; Fast mode and regional processing have separate premiums.

MiniMax M3.1 Flash

MiniMax's multimodal reasoning model for coding, tool use, and long-running agents, with a 1,000,000-token context and text, image, and video input. Available through the MiniMax API; downloadable weights are announced but not yet released. OpenRouter currently offers a free endpoint, subject to its availability and usage limits.

Recursive Flint

Recursive's open-weight mixture-of-experts language model for coding agents, with 39.8B total and approximately 2.5B active parameters per token. Supports text generation, reasoning and multi-turn tool use under Apache 2.0. The released checkpoint configures 262,144 tokens of combined input and output; the developer advertises up to 512K but recommends staying within 256K. OpenRouter currently offers a free endpoint subject to availability and usage limits.

Mercury 3

Inception's diffusion-based language model for low-latency reasoning, coding, and tool-using agents, with a 1,000,000-token context. Current launch pricing is $0.10 input and $0.35 output per million tokens, with $0.01 cached input. Announced standard pricing is $0.20 input and $0.70 output; no end date for the launch discount is published.