CharXiv Reasoning without tools
Tool-free variant of CharXiv that isolates raw visual reasoning ability without code execution or tool augmentation.
14 models
Higher is better.
- 1Claude Mythos 5Anthropic88.9
- 2Qwen3.8 MaxAlibaba88.4
- 3Gemini 3.8 FlashGoogle86.2
- 4Kimi K3Moonshot AI84.8
- 5Qwen3.8-Flash-NextAlibaba84.6
- 6Gemini 3.7 FlashGoogle84.5
- 7Qwen3.8-27BAlibaba83.7
- 8Qwen3.8-Omni-FlashAlibaba83.5
- 9dots3-note PreviewDots Studio83.1
- 10Claude Opus 4.7 (Adaptive)Anthropic82.1
- 11Claude Opus 4.8Anthropic80.5
- 12InklingThinking Machines Lab78.1
- 13Inkling-SmallThinking Machines Lab77.4
- 14Claude Sonnet 5Anthropic77.0
About this benchmark
- Metric
- Chart understanding without tools
- Tasks
- Scientific chart reasoning (tool-free)
- Authors
- CharXiv authors (2024)
Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.