Inference
EN

CharXiv Reasoning without tools

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

ModelsModels with a result.17
Top resultBest result on this benchmark.89.0%Kimi K3.1 Pro
Top-3 spreadPoints from first to third.0.6 pts
YearYear of release.—

Result and price

13 of 17 results

Results

17 results
#ModelResultThe score from the source.
01Kimi K3.1 Pro89.0%
02Claude Mythos 588.9%
03Qwen3.8 Max88.4%
04Gemini 3.8 Flash86.2%
05Inkling 2 Small85.6%
06Kimi K384.8%
07Qwen3.8-Flash-Next84.6%
08Gemini 3.7 Flash84.5%
09Qwen3.8-27B83.7%
10Qwen3.8-Omni-Flash83.5%
11dots3-note Preview83.1%
12Claude Opus 4.7 (Adaptive)82.1%
13Claude Opus 4.880.5%
14Inkling78.1%
15Inkling-Small77.4%
16Claude Sonnet 577.0%
17LFM3-VL-2B70.2%