kimi-k3

MiniMax-M3.1-Flash is a multimodal model from MiniMax. Compared with M3, it delivers significant improvements across multiple benchmarks, particularly in complex agentic tasks, coding, and cybersecurity. It is built for complex, multi-step co-work automation, understanding real codebases, fixing bugs, and building features, and visual understanding and creation, including frontend and presentation creation. Open weights are planned.

MiniMax-M3.1-Flash is a multimodal model from MiniMax. Compared with M3, it delivers significant improvements across multiple benchmarks, particularly in complex agentic tasks, coding, and cybersecurity. It is built for complex, multi-step co-work automation, understanding real codebases, fixing bugs, and building features, and visual understanding and creation, including frontend and presentation creation. Open weights are planned.

Currently powered by DeepSeek-V4.1-Flash-1017, with native multimodal visual understanding and support for both thinking and non-thinking modes. Suitable for complex reasoning, coding, long-document analysis, visual understanding, and tool calling.

A next-generation omni-modal large model with native unified understanding of text, vision, and audio. It features a 1M-token context window and strong overall multimodal capabilities, with significant improvements in task completion for coding and agentic scenarios.

Seed-Evolving is a model in the Seed series designed for agentic workflows, multimodal understanding, and software development. This update further strengthens complex task planning, visual reasoning, tool execution, and result verification, enabling the model to better understand images, videos, and real software environments, and to more reliably complete the full workflow from information gathering to final delivery. Use the unified Model ID **`doubao-seed-evolving`** to continuously access the latest model capabilities without needing to track or switch between model versions.