ATRIUM search → argument graph

FactArticle

Artificial Analysis ranks DeepSeek-V4-Flash-0731 ahead of MiniMax M3, a 428-billion-parameter model.

Notes that Artificial Analysis places DeepSeek-V4-Flash-0731 above MiniMax M3, a larger 428B-parameter model. ✦ AI generated

Author · Simon Willison's Weblog · 2026-07-31 · original ↗

Share card ↓Export .md Export .txt

Artificial Analysis rank it ahead of MiniMax M3 - a 428B model.

Read full article ↗excerpt · fair-use quotation

▦ Artificial Analysis ▦ MiniMax ✳ MiniMax M3

Around this claim

Evidence · 1

DeepSeek-V4-Flash-0731, the latest release in DeepSeek's V4 family, has substantially enhanced agentic capabilities and, at 304 billion parameters (167GB on Hugging Face), punches well above its weight.Author · Simon Willison's Weblog · conf 85%

This moment responds to

supports → At $0.14 per million input tokens and $0.27 per million output tokens, DeepSeek-V4-Flash-0731 may currently be the best value-per-intelligence model available, performing very well on the Intelligence Index versus cost.Author · Simon Willison's Weblog