AI Office

The new Opus for complex coding and professional work, with lower standard token prices.

Updated 2026-10-09

A balanced coding and agent model, faster than Sonnet 5 for everyday workflows.

Updated 2026-10-09

Near-Astra capability at a lower API cost for complex coding, computer use, and professional work.

Updated 2026-10-09

A new fast small model for summaries, classification, customer support, and coding subagents.

Updated 2026-10-09

DeepSeek's new native-vision model for reasoning and coding, served as deepseek-flash.

Updated 2026-10-09

Doubao's continuously updated model endpoint for long-running agents, coding, and multimodal understanding.

Updated 2026-10-09

Google's recommended general Flash model for text, images, video, audio, and PDF inputs.

Updated 2026-10-09

Z.ai's coding and agent flagship with million-token context and text-only inputs.

Updated 2026-10-09

OpenAI's high-capability GPT-6 model for demanding reasoning and long-running work.

Updated 2026-10-09

The new Grok workhorse for coding and knowledge work, with vision, configurable reasoning, and tools.

Updated 2026-10-09

Kimi's new multimodal open-weight flagship for long documents and coding agents.

Updated 2026-10-09

Qwen's cloud Max model with million-token context and vision for documents and complex agents.

Updated 2026-10-09

Google's new image generation and editing model for visual content and text-rich graphics.

Updated 2026-10-09

Z.ai's released fast multimodal weights for vision and coding agents under MIT.

Updated 2026-10-09

The cost-efficient GPT-6 option for focused tasks, summaries, and high-volume workloads.

Updated 2026-10-09

Mistral's new multimodal flagship is in public API preview, with weights planned for month-end.

Updated 2026-10-09

Meta's new coding and agent model for collaborative long-running work on Muse Code and its Model API.

Updated 2026-10-09

A cost-efficient hosted Flash model with million-token context and built-in tools.

Updated 2026-10-09

ByteDance's new joint audio-video model for 30-second storytelling, reference control, and editing.

Updated 2026-10-09

ByteDance's new image model for text-rich infographics, annotated editing, and multilingual graphics.

Updated 2026-10-09

The production V4 Pro revision for text reasoning and coding, without vision input.

Updated 2026-10-09

Doubao's new general model family for office work, coding delivery, and visual understanding.

Updated 2026-10-09

Qwen's new unified image model for generation, editing, and native transparency under a research license.

Updated 2026-10-09

An open 27B vision-language model for self-hosting, with optional context extension.

Updated 2026-10-09