Atlas

Benchmarks

← All benchmarks

Code Migration — Code Quality

Code · 2026-06-16

Code-quality rubric score (0–5) sub-track of the Code Migration benchmark. Split out because it is a 0–5 rubric, not a 0–100 pass rate.

Top models (higher is better)

ModelScore
Sonnet 54.9
Claude Haiku 5.54.8
Claude Opus 54.8
Opus 4.84.7
Claude Sonnet 5.54.6
Opus 4.74.6
GLM-5.34.5
GPT-6.1 Sol4.5
GPT-6 Astra4.5
Hy4 Preview4.4
DeepSeek V4.1 Flash4.4
Step 5 Preview4.4
GPT-6 Sol4.4
GLM-5.3 Flash4.4
Muse Spark 1.34.3
Qwen3.8-Max4.3
Muse Wonder Full4.2
MiMo-V2.6-Pro4.2
Qwen3.8 27B4.2
Mistral Large 4 Preview4.1
Grok 4.74.1
MiMo-V2.6-Flash4.1
Grok 4.54.1
Grok 4.64.1
Muse Wonder4.1
DeepSeek V4 Pro 08134.0
Sonnet 4.64.0
DeepSeek V4.1 Flash 10174.0
GLM-5.24.0
Ember-14.0
DeepSeek-V4-Flash-07313.9
GPT-5.6 Terra3.9
Kimi K33.9
GPT-5.6 Sol3.8
GPT-6 Luna3.8
Gemini 3.8 Flash3.7
MiniMax M33.7
Gemini 3.7 Flash3.7
GPT-5.4 Nano3.7
GPT-5.43.6
Loading Atlas data…