# Zhao Pengyu

> Zhao Pengyu is executive director and large language model research leader at MiniMax. MiniMax API, MiniMax Code, MiniMax-M2.5, MiniMax-M2.7, MiniMax-M3.1-Flash, MiniMax Video API and MiniMax Voice Clone and Voice Design on Anchor Terminal, with grades, an about and public profiles.

- Canonical: https://www.anchorterminal.com/people/zhao-pengyu/
- Markdown: https://www.anchorterminal.com/people/zhao-pengyu/index.md (~1,700 tokens)
- Slim: https://www.anchorterminal.com/people/zhao-pengyu/index.min.md (~1,480 tokens, same facts, less prose, for token-sensitive contexts)
- JSON: https://www.anchorterminal.com/people/zhao-pengyu/index.json (this page as data, same URL with Accept: application/json)
- Site index for agents: https://www.anchorterminal.com/llms.txt (full text: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- MCP: https://www.anchorterminal.com/mcp (the whole catalogue as tools, no key)
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)
- Updated: 2026-10-28

Executive director and large language model research leader, MiniMax.

Zhao Pengyu leads research and engineering on large language models at MiniMax and has been an executive director since June 2025. He joined as a natural language processing researcher in August 2023, after three years on recommendation algorithms at Hulu in Beijing, and studied computer science at Peking University.

## What Zhao builds

- [MiniMax API](https://www.anchorterminal.com/tools/minimax-api.md), graded C, 55.3/100, in Model APIs & inference. MiniMax's API for its M-series text models, M3.1 Flash and M3, in the Anthropic Messages, OpenAI Chat Completions and OpenAI Responses formats, with 1M-token context and automatic prompt caching. Executive director and large language model research leader, MiniMax, researched 2026-10-28.
- [MiniMax Code](https://www.anchorterminal.com/tools/minimax-code.md), graded B, 67.7/100, in Agent harnesses. MiniMax's coding agent, as an open-source terminal CLI (`mcode`) and a closed desktop app. The CLI runs MiniMax models through M Plan or an API key, or any compatible model, headless with `mcode exec` or as an ACP server. Executive director and large language model research leader, MiniMax, researched 2026-10-28.
- [MiniMax-M2.5](https://www.anchorterminal.com/tools/minimax-m2-5.md), graded C, 57.5/100, in Open-weight models. MiniMax-M2.5 is MiniMax's open-weight mixture-of-experts model for coding and agent work, about 229B parameters in FP8 with a 196,608-token context and thinking always on, published on Hugging Face on 12 February 2026. Executive director and large language model research leader, MiniMax, researched 2026-10-28.
- [MiniMax-M2.7](https://www.anchorterminal.com/tools/minimax-m2-7.md), graded C, 56.1/100, in Open-weight models. MiniMax-M2.7 is MiniMax's open-weight mixture-of-experts model for coding, agent harnesses and office work, about 229B parameters in FP8 with thinking always on, published on Hugging Face on 9 April 2026 under a non-commercial licence. Executive director and large language model research leader, MiniMax, researched 2026-10-28.
- [MiniMax-M3.1-Flash](https://www.anchorterminal.com/tools/minimax-m3-1-flash.md), graded D, 50.5/100, in Open-weight models. MiniMax-M3.1-Flash is MiniMax's open-weight mixture-of-experts model for coding, agent work, computer use and cybersecurity, about 428B parameters with 23B active, taking text, images and video with up to 1M tokens of context. Executive director and large language model research leader, MiniMax, researched 2026-10-28.
- [MiniMax Video API](https://www.anchorterminal.com/tools/minimax-video.md), graded C, 61.1/100, in Video generation. MiniMax's API for generating video. Executive director and large language model research leader, MiniMax, researched 2026-10-28.
- [MiniMax Voice Clone and Voice Design](https://www.anchorterminal.com/tools/minimax-voice-cloning.md), graded D, 53.2/100, in Voice cloning & custom voices. MiniMax's API clones a voice from a 10 second to 5 minute audio sample and designs voices from a text description. The resulting voice IDs work with its speech-2.8 text-to-speech models. Executive director and large language model research leader, MiniMax, researched 2026-10-28.

## Works with

- [Xunhao Lai](https://www.anchorterminal.com/people/xunhao-lai/), First author of MiniMax Sparse Attention
- [Yan Junjie](https://www.anchorterminal.com/people/yan-junjie/), Founder, chairman, CEO and CTO, MiniMax
- [Yeyi Yun](https://www.anchorterminal.com/people/yeyi-yun/), Co-founder, executive director and COO, MiniMax
- [Zhou Yucong](https://www.anchorterminal.com/people/zhou-yucong/), Co-founder, executive director and visual model research leader, MiniMax

## Questions

### Who is Zhao Pengyu?

Zhao Pengyu leads research and engineering on large language models at MiniMax and has been an executive director since June 2025. He joined as a natural language processing researcher in August 2023, after three years on recommendation algorithms at Hulu in Beijing, and studied computer science at Peking University.

### What does MiniMax make?

Anchor Terminal lists MiniMax API, MiniMax Code, MiniMax-M2.5, MiniMax-M2.7, MiniMax-M3.1-Flash, MiniMax Video API and MiniMax Voice Clone and Voice Design from MiniMax. MiniMax's API for its M-series text models, M3.1 Flash and M3, in the Anthropic Messages, OpenAI Chat Completions and OpenAI Responses formats, with 1M-token context and automatic prompt caching.

### How are Zhao's tools graded?

MiniMax API is graded C, 55.3 out of 100; MiniMax Code is graded B, 67.7 out of 100; MiniMax-M2.5 is graded C, 57.5 out of 100; MiniMax-M2.7 is graded C, 56.1 out of 100; MiniMax-M3.1-Flash is graded D, 50.5 out of 100; MiniMax Video API is graded C, 61.1 out of 100; MiniMax Voice Clone and Voice Design is graded D, 53.2 out of 100, on Anchor Terminal's benchmark, which scores every listing against the same checklist.

## Sources

- https://cn.linkedin.com/in/yeyi-yun-ba051b220
- https://www.minimax.io/news/minimax-announces-first-half-2026-financial-results-1787744160
- https://www.minimax.io
- https://agent.minimax.io/docs/code/welcome
- https://www.minimax.io/about
- https://agent.minimax.io/docs/code/help/feedback
- https://platform.minimax.io/docs/faq/contact-us
- https://agent.minimax.io/docs/miniapp/overview
- https://www.minimax.io/blog/minimax-code-mini-app-1792650729
- https://agent.minimax.io/doc/en/privacy-policy.html
- https://huggingface.co/MiniMaxAI/MiniMax-M3.1-Flash
- https://arxiv.org/abs/2606.13392
- https://xunhaolai.github.io/

These are the people building the tools we list, named for their public role in the company or project behind a listing, from public pages, with the date it was researched. They don't work for Anchor Terminal (our own team is on the about page, /about/#author). We show public roles, public profiles and what the listing pages show, nothing private. To correct something, send an about or a photo, or have a page removed, write to privacy@anchorterminal.com.

---

Query this catalogue directly over MCP: https://www.anchorterminal.com/mcp (Streamable HTTP, no key). How: https://www.anchorterminal.com/agents/#mcp