FeynmanBench - Topological Connectivity: leaderboard

FeynmanBench (2026) gives a multimodal model a machine-generated Feynman diagram (Standard Model QED, electroweak or QCD, tree level to two loops) with minimal labelling conventions and asks it to recover the full physical content of the diagram. Checkpoint 3: a vertex connection graph isomorphic to the diagram, judged by a three-model LLM panel; pass rate in percent, higher is better.

Metric: CP3 pass rate (%), a vertex connection graph isomorphic to the diagram, judged by a three-model LLM panel by majority vote, on FeynmanBench's Feynman-diagram tasks (a rendered diagram plus minimal conventions; Standard Model QED, electroweak and QCD diagrams from tree level to two loops), API models at temperature 0, mean of five runs; four models ran on subsets of 165 to 742 items, the others on about 2,000; higher is better. Source: arxiv.org. Saturation forecast: Around November 2027. 17 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Gemini 3.1 Pro (Preview)33.6#99
2Claude Sonnet 4.619.2#173
3Gemini 3 Flash16.8#175
4Qwen 3.5 Plus15.9#242
5Qwen 3.6 Plus15.2#238
6Seed 2.0 Pro13.3#222
7GLM-5V Turbo10.5#360
8Claude Opus 4.75.9#85
9Claude Opus 4.65.7#113
10Claude Opus 4.53.5#160
11GPT-5.23.3#185
12GPT-5.13.1#240
13GPT-5.43#135
14Kimi K2.51.4#266
15Claude Sonnet 4.51.1#261

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=feynmanbench-topological-connectivity · How It Works · Data refreshed daily, snapshot 2026-11-21.