Calibration

WideResearch → HLE w/ tools

Name: benchgap: LLM benchmark scores, measured and estimated
Creator: benchgap
License: https://creativecommons.org/licenses/by-nc/4.0/

HLE w/ tools is estimated from WideResearch with a linear curve fitted on 7 models measured on both: y = 1.1150·x + -0.3544, R² = 0.84, cross-validated error 5.3 pp. It is used for 0 estimates.

Fitted on

Model	WideResearch	HLE w/ tools
dots3-note Preview	78.9%	52.6%
GLM-5	69.8%	50.4%
Hy4 preview	83.9%	55.4%
Ornith-1.5-35B-A3B	67.8%	33.4%
Ornith-1.5-397B	80.8%	56.1%
Ornith-1.5-9B	59.5%	30.5%
Qwen3.8 Max	81.9%	56.2%