# CharXiv w/o tools leaderboard

Multimodal & Grounded benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelscale.dev/leaderboards/benchmarks/multimodal-grounded-charxiv-no-tools

CharXiv Reasoning without tools. Every model the catalog carries a published CharXiv w/o tools value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Multimodal & Grounded |
| Measure | Chart understanding without tools |
| Tasks | Scientific chart reasoning (tool-free) |
| Difficulty | Scientific visualization reasoning |
| Published by | [CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs](https://charxiv.github.io/) |

Tool-free variant of CharXiv that isolates raw visual reasoning ability without code execution or tool augmentation.

## CharXiv w/o tools ranking

19 models with a published CharXiv w/o tools 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 without tools | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [Kimi K3.1 Pro](https://modelscale.dev/models/kimi-3-1-pro) | Moonshot AI | 89 | Observed 2026-11-21 · source benchlm:models |
| 2 | [Claude Mythos 5](https://modelscale.dev/models/claude-mythos-5) | Anthropic | 88.9 | Observed 2026-11-21 · source benchlm:models |
| 3 | [Qwen3.8 Max](https://modelscale.dev/models/qwen3-8-max) | Alibaba | 88.4 | Observed 2026-11-21 · source benchlm:models |
| 4 | [Gemini 3.8 Flash](https://modelscale.dev/models/gemini-3-8-flash) | Google | 86.2 | Observed 2026-11-21 · source benchlm:models |
| 5 | [Inkling 2 Small](https://modelscale.dev/models/inkling-2-small) | Thinking Machines Lab | 85.6 | Observed 2026-11-21 · source benchlm:models |
| 6 | [Kimi K3](https://modelscale.dev/models/kimi-3) | Moonshot AI | 84.8 | Observed 2026-11-21 · source benchlm:models |
| 7 | [Qwen3.8-Flash-Next](https://modelscale.dev/models/qwen3-8-flash-next) | Alibaba | 84.6 | Observed 2026-11-21 · source benchlm:models |
| 8 | [Gemini 3.7 Flash](https://modelscale.dev/models/gemini-3-7-flash) | Google | 84.5 | Observed 2026-11-21 · source benchlm:models |
| 9 | [Qwen3.8-27B](https://modelscale.dev/models/qwen3-8-27b) | Alibaba | 83.7 | Observed 2026-11-21 · source benchlm:models |
| 10 | [Qwen3.8-Omni-Flash](https://modelscale.dev/models/qwen3-8-omni-flash) | Alibaba | 83.5 | Observed 2026-11-21 · source benchlm:models |
| 11 | [dots3-note Preview](https://modelscale.dev/models/dots3-note-preview) | Dots Studio | 83.1 | Observed 2026-11-21 · source benchlm:models |
| 12 | [Claude Opus 4.7](https://modelscale.dev/models/claude-opus-4-7) | Anthropic | 82.1 | Observed 2026-11-21 · source benchlm:models |
| 12 | [Claude Opus 4.7 (Adaptive)](https://modelscale.dev/models/claude-opus-4-7-max) | Anthropic | 82.1 | Last good 2026-11-18 · source benchlm:models |
| 14 | [Inkling 2 Mini](https://modelscale.dev/models/inkling-2-mini) | Thinking Machines Lab | 80.6 | Observed 2026-11-21 · source benchlm:models |
| 15 | [Claude Opus 4.8](https://modelscale.dev/models/claude-opus-4-8) | Anthropic | 80.5 | Observed 2026-11-21 · source benchlm:models |
| 16 | [Inkling](https://modelscale.dev/models/inkling) | Thinking Machines Lab | 78.1 | Observed 2026-11-21 · source benchlm:models |
| 17 | [Inkling-Small](https://modelscale.dev/models/inkling-small) | Thinking Machines Lab | 77.4 | Observed 2026-11-21 · source benchlm:models |
| 18 | [Claude Sonnet 5](https://modelscale.dev/models/claude-sonnet-5) | Anthropic | 77 | Observed 2026-11-21 · source benchlm:models |
| 19 | [LFM3-VL-2B](https://modelscale.dev/models/lfm3-vl-2b) | LiquidAI | 70.2 | Observed 2026-11-21 · source benchlm:models |

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
