benchgap
Calibration

ResearchClawBench → Gert Labs

Gert Labs is estimated from ResearchClawBench with a Michaelis–Menten + offset curve fitted on 16 models measured on both: y = 0.0000 + 2.0000·x / (0.42400 + x), R² = 0.41, cross-validated error 8.5 pp. It is used for 1 estimate.

Estimated modelResearchClawBenchGert LabsSource
Grok 4.113.5%48.3%estimated ± 8.5 pp, low confidence

Fitted on

ModelResearchClawBenchGert Labs
Claude Opus 4.619.9%61.9%
Claude Opus 4.720.7%65.6%
Claude Opus 4.821.1%73.0%
Gemini 3.1 Pro13.3%56.9%
Gemini 3.5 Flash18.0%61.9%
GLM-5.118.2%60.1%
GPT-5.415.3%64.9%
GPT-5.517.0%72.9%
Grok 4.312.4%43.9%
Kimi K2.618.0%56.8%
Kimi K2.514.0%45.9%
MiMo-V2.516.9%46.9%
MiMo-V2-Pro15.3%36.7%
Qwen3.5 397B14.2%46.8%
Qwen3.6 Plus18.0%50.6%
Qwen3.7 Max18.7%64.3%