DeepSeek V4.1 Flash tops open-model usage among research agents, index data shows
AkariAsai · x · 2026-10-30
Trends from the Artificial Analysis Intelligence Index v5.0 and OpenResearch usage data on research-agent users:
- Terminal-Bench-Science debuts in the index (15% of total), with the Science category up-weighted from 20% to 25%.
- Anthropic and OpenAI models dominate usage, taking 6 of the 7 most popular slots; Opus 5.5 leads both intelligence and usage (122.5B tokens).
- DeepSeek V4.1 Flash is the most popular open and Chinese model by a wide margin (20.6B tokens, rank 4), favored by researchers for its cheap intelligence.
- Clear popularity mismatch: GPT-6 Luna's usage far outpaces its intelligence rank (+9), while Haiku 5.5 and Grok 4.7 are most underused (-7).
- Google's latest Gemini 4 models have yet to make meaningful inroads in research usage.
The attached chart juxtaposes intelligence rank vs. usage rank for 18 models, highlighting the gap between benchmark scores and real-world adoption.
More from Models
- Kimi K3.1 Pro leak teases open weights by Nov 4, 2026 as progress index reportedly tops Fable 5 — teortaxesTex · 2026-10-30
- Users complain Opus 5.5 writes 'illegible rushed technical speak' and won't fix it — msalbergo · 2026-10-30
- Dev claims Gemini Argon makes 'researcher-level jumps' other models miss as 'carbon' looms — vitaliychiley · 2026-10-30
- Meta's Alexandr Wang touts Muse: 6.7x more concise than ChatGPT at answering the same question — alexandr_wang · 2026-10-30
- OpenAI's subsidy shift rewrites the score-vs-cost Pareto frontier, analysis finds — RexDouglass · 2026-10-30
- Phonic tops VAmoS benchmark for agentic speech-to-speech models — graceisford · 2026-10-30