benchgap
Vision & documents

ZeroBench w/ Python leaderboard

As of 2026-11-21, the highest measured score on ZeroBench w/ Python is 63.0% by Kimi K3.1 Pro. 4 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#ModelScoreSource
1Kimi K3.1 Pro63.0%measured
2Glimpse56.4%estimated ± 7.0 pp, low confidence
3Hy455.4%measured
4Inkling 2 Small53.3%estimated ± 7.0 pp, low confidence
5DeepSeek V4.1 Flash 101751.0%measured
6Qwen3.8 Max49.0%measured
7MiniMax M3.1 Flash48.1%estimated ± 7.0 pp, low confidence
8Hy4 flash45.8%measured
9Kimi K341.0%measured
10Qwen3.8-27B38.0%estimated ± 7.0 pp, low confidence
11Inkling 2 Mini34.0%measured

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General