Multimodal & Grounded benchmark
CharXiv leaderboard
CharXiv Reasoning. Every model the catalog carries a published CharXiv value for, ranked by that value.
A scientific chart reasoning benchmark that tests whether models can understand, interpret, and reason about complex scientific visualizations including plots, diagrams, and data charts.
CharXiv ranking
39 models with a published CharXiv value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Chart understanding and reasoning |
|---|---|---|---|
| 1 |  Claude Mythos 5 | Anthropic | 93.5 |
| 1 |  Qwen3.8 Max | Alibaba | 93.5 |
| 3 | Qwen3.8-Omni-Flash | Alibaba | 91.4 |
| 4 |  Kimi K3 | Moonshot AI | 91.3 |
| 5 | Claude Opus 4.7 (Adaptive) | Anthropic | 91 |
| 6 | Qwen3.8-Flash-Next | Alibaba | 90.6 |
| 7 | Qwen3.8-27B | Alibaba | 90.2 |
| 8 | Claude Opus 4.8 | Anthropic | 89.9 |
| 9 |  GLM-5.3-Flash | Z.AI | 89.4 |
| 10 |  Gemini 3.7 Flash | 88.7 | |
| 11 |  Muse Spark 1.1 | Meta | 88.4 |
| 12 | Claude Sonnet 5 | Anthropic | 88.3 |
| 13 | SASakana Fugu-Ultra | Sakana AI | 86.6 |
| 14 | Muse Spark | Meta | 86.4 |
| 15 | Qwen3.7 Plus | Alibaba | 85.9 |
| 16 |  Seed 2.1 Pro | ByteDance | 85.4 |
| 17 | SASakana Fugu | Sakana AI | 85.1 |
| 18 | Gemini 3.5 Flash | 84.2 | |
| 19 |  GPT-5.4 | OpenAI | 82.8 |
| 20 | Seed 2.1 Turbo | ByteDance | 82.5 |
| 21 | GPT-5.2 | OpenAI | 82.1 |
| 22 | TMInkling | Thinking Machines Lab | 82 |
| 23 | Qwen3.6 Plus | Alibaba | 81.5 |
| 24 | Gemini 3 Pro | 81.4 | |
| 25 | TMInkling-Small | Thinking Machines Lab | 81.3 |
| 26 |  MiMo-V2.5 | Xiaomi | 81 |
| 27 | Qwen3.5 397B | Alibaba | 80.8 |
| 28 | Kimi K2.6 | Moonshot AI | 80.4 |
| 29 | Gemini 3.1 Pro | 80.2 | |
| 30 | Muse Glimmer 30B | Meta | 78.8 |
| 31 | Qwen3.6-27B | Alibaba | 78.4 |
| 32 | Qwen3.6-35B-A3B | Alibaba | 78 |
| 33 | Claude Sonnet 4.6 | Anthropic | 77.4 |
| 34 | Qwen3.5-122B-A10B | Alibaba | 77.2 |
| 35 |  Nemotron 3 Nano Omni 30B A3B | NVIDIA | 76.25 |
| 36 | Gemini 3.1 Flash-Lite | 73.2 | |
| 37 | Claude Opus 4.5 | Anthropic | 68.5 |
| 38 |  Grok 4.20 | xAI | 60.9 |
| 39 |  Command A+ | Cohere | 52.7 |
Evidence key: Observed
Rows are ordered by the value CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.