LLM Stats (SWE Atlas - Codebase QnA): leaderboard

SWE Atlas Codebase QnA evaluates repository-level code comprehension: answering questions about code structure, behavior and intent across real software projects. LLM Stats mirrors reported results as percentages.

Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around July 2027. 8 models tracked. Updated 13 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Mistral Large 4 Preview59.4#206
2Muse Spark 1.359.4#50
3Muse Spark Open58#82
4SWE-2 Flash50#73
5K2 Horizon 375B A23BHigh48.4#291
6Laguna S 2.146.2#768
7MiniMax-M337.9#290
8Beam 1.031.8#215

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=llm-stats-swe-atlas-codebase-qna · How It Works · Data refreshed daily, snapshot 2026-11-22.