Benchgen

CharXiv Reasoning — Results

RankModelScore
1kimi-k391.3
2qwen3-8-flash-next90.6
3qwen3-8-27b90.2
4fugu-ultra86.6
5fugu85.1
6minimax-m3-1-flash82.5
7inkling82
8muse-glimmer78.8
9lfm3-vl-2b70.2
10kimi-k30.913
11claude-opus-4-70.91
12claude-opus-4-80.899
13muse-spark0.884
14claude-sonnet-50.883
15kimi-k2-60.867
16seed-2-1-pro0.864
17qwen3-7-plus0.859
18seed-2-1-turbo0.836
19gpt-5-20.821
20gpt-5-50.816
21qwen3-6-plus0.815
22gemini-3-pro0.814
23gpt-50.811
24gemini-3-flash0.803
25qwen3-5-27b0.795

CharXiv Reasoning — Phases

1 phaseActive

Phase 1·

CharXiv Reasoning

Jun 22, 2026
Current