Agents
Eval
Train
New Evaluation
Evaluations
Results
Datasets
Models
Recents
View All
Your evaluations
Sign in to run benchmarks and follow your results here.
Website
Docs
Sign In
Toggle Sidebar
Benchgen
Overview
Leaderboard
Evaluations
Phases
CharXiv Reasoning (No Tools) — Results
Rank
Model
Score
1
claude-fable-5
86.5
2
kimi-k3
84.8
3
gpt-5-6-sol
84.7
4
kimi-k2-6
80.4
5
gemini-3-1-pro
80.2
6
inkling
78.1
7
kimi-k2-5
77.5
CharXiv Reasoning (No Tools) — Phases
1 phase
Active
Evaluate
Overview
Phases
1
Leaderboard
Evaluations
Evaluate
Phase 1
·
CharXiv Reasoning (No Tools)
Jul 16, 2026
Current
Evaluate