DeepSeek V4.1F tops FrontierMath Tier 4 among Chinese models at 46%, ECI estimated ~158
teortaxesTex · x · 2026-10-26
teortaxesTex estimates DeepSeek V4.1F-1017's capability: it becomes the strongest Chinese model on FrontierMath Tier 4 v2 at 46%, matching Qwen 3.8 Max (Aug 2) and Haiku 5.5, up sharply from 24–27% for prior DeepSeek versions — a sign DeepSeek is hill-climbing math again.
Key points from the ECI extrapolation:
- Official ARC scores: ARC-AGI-2 at 78.1% (predicted 77%), ARC-AGI-1 at 94.5%; updated Epoch Capability Index estimate ≈158.2 (range 156.9–160.7).
- A regression across 19 models on the Artificial Analysis index (45.1, +6.1 over V4.1 Flash) yields ECI ≈ AA×0.634 + 132.8 − 3.95 open-weights adjustment; multiple methods span 156.7–159.8.
- Open-weights models score 4 ECI points lower at the same AA index; no independent third-party results exist yet for other ECI benchmarks like SWE-bench Verified or Terminal-Bench 2.0.
Related event: DeepSeek V4.1F Tops FrontierMath Tier 4 as Best Chinese Model(2 posts)→
More from Models
- Lyra neuron explorer: steering Qwen 3.5 4B activations turns it into a 'money-making' model — rudrank · 2026-10-26
- Harness switch reveals underrated Gemini 4 Argon, now second only to Opus 5.5 — banteg · 2026-10-26
- Leak: Kimi K3.1 Flash reportedly in training, DeepSeek 4.1 Pro still 'grinding' — teortaxesTex · 2026-10-26
- Editing just 2 neurons instantly "flips" Qwen 3.5 4B, netizens joke — jaivinwylde · 2026-10-26
- ChatGPT Pro user with 95% quota left hit with "capabilities reduced" notice — Shozab_haxor · 2026-10-26
- Google auto-converts one-off searches into AI Mode conversations, users frustrated — Status_Technician364 · 2026-10-26