DeepSeek V4.1 Flash 1017 hits 45 on AA Intelligence Index, tying #2 open-weights
ArtificialAnlys · x · 2026-10-20
DeepSeek released V4.1 Flash 1017, an update to its flagship that scores 45 on the Artificial Analysis Intelligence Index (+6), tying GLM-5.3 as the #2 open-weights model, just behind MiMo-V2.6-Pro (46).
Key evals
- GDPval-AA v2.1 Elo up 146 points (1600 → 1746); AA-Briefcase v1.1 up 150 points (1420 → 1570)
- AA-Omniscience non-hallucination rate jumped from 4% to 33%, pushing the index from -5 to +5
- CritPt physics reasoning improved from 14% to 24%
- Uses 89K output tokens per Intelligence Index task
Specs & pricing
- 552B total params (16B active), 1M-token context, MIT license
- Unchanged pricing: $0.30/$1.20 per 1M input/output tokens, $0.006 cache hits (98% discount), half-price off-peak
- $0.26 cost per task, in the most attractive capability-vs-cost quadrant
Available via DeepSeek's first-party API plus Fireworks, Baseten, DeepInfra, Cloudflare and others.
Related event: DeepSeek V4.1-Flash-1017: 'Minor Update' Tops Open-Weight Leaderboards(12 posts)→
More from Models
- lintool's Gaggle Matches SFT'd Qwen With Just a Handful of GPUs on nanochat Backbone — lintool · 2026-10-20
- Report: Liang Wenfeng planned independent DeepSeek compute chips before Huawei persuaded him otherwise — teortaxesTex · 2026-10-20
- Smallest AI's full voice stack hits 26.8% completion at $0.048/call on 300 hard calls — auto_grad_ · 2026-10-20
- NVIDIA's Dory1.0-4B Looped Model Lets You Pick Inference Depth at Runtime — lmoroney · 2026-10-20
- OpenAI dropped 722 AI-generated math papers in a day; can humans read them? — songyoupeng · 2026-10-20
- JevBench adds 34 decision models in four days; Nimble-v4-max price cut to $0.04/1M tokens — airesearch12 · 2026-10-20