benchgap
Calibration

Vibe Code Bench → SWE-bench Pro

SWE-bench Pro is estimated from Vibe Code Bench with a linear curve fitted on 14 models measured on both: y = 0.0873·x + 0.5259, R² = 0.34, cross-validated error 2.9 pp. It is used for 0 estimates.

Fitted on

ModelVibe Code BenchSWE-bench Pro
Claude Opus 4.657.6%53.4%
Claude Opus 4.771.0%64.3%
DeepSeek V4 Pro49.9%55.4%
Gemini 3.5 Flash48.7%55.1%
GLM-5.131.5%58.4%
GPT-5.253.5%55.6%
GPT-5.3 Codex61.8%56.8%
GPT-5.467.4%57.7%
GPT-5.569.8%58.6%
Grok 4.204.1%51.8%
Kimi K2.637.9%58.6%
MiniMax M2.727.0%56.2%
Muse Spark19.7%52.4%
Qwen3.6 Plus25.6%56.6%