DeepSeek V4.1F Tops FrontierMath Tier 4 as Best Chinese Model

According to Epoch's FrontierMath Tier 4 (v2) benchmark data, DeepSeek V4.1F-1017 scored 46%, making it the highest-ranked Chinese model on the leaderboard, surpassing competitors such as Qwen.

2026-10-26 ~ 2026-10-26 · 2 related posts

• DeepSeek V4.1F tops Chinese models on FrontierMath Tier 4 at 46% — teortaxesTex · 2026-10-26

1 near-duplicate retellings: teortaxesTex