Atlas

Benchmarks

← All benchmarks

Senior SWE-Bench Basic Solve Rate

Code · 2026-06-30

Senior SWE-Bench basic solve rate counts tasks whose functional requirements are satisfied without applying the benchmark's stricter taste and engineering-quality criteria.

Top models (higher is better)

ModelScore
Claude Opus 562.1
Claude Fable 553.7
GPT-5.6 Sol53.7
GPT-5.552.6
Sonnet 550.5
Opus 4.746.8
Opus 4.844.2
GPT-5.441.1
Kimi K340.4
GPT-5.6 Terra36.8
GLM-5.235.8
Gemini 3.5 Flash22.1
GPT-5.6 Luna20.3
Gemini 3.1 Pro Preview9.5
Loading Atlas data…