Code Migration v1.1

Reimplementing real-world programs (CLI tools, COBOL) in another language. Run by Vals AI.

CategoryCoding
MetricAccuracy (%)
SourceVals AI (evaluator)
LicenseNone stated
Benchmark pagewww.vals.ai
StatusScored
Weight1.0 within Coding
FitDiscrimination 1.57, noise 0.32

Results

#ModelProviderResultDate
1 Claude Sonnet 5.5 Max Anthropic 86.9% 2026-11-09
2 Claude Fable 5.2 High Anthropic 83.3% 2026-11-09
3 Claude Opus 5.5 Max Anthropic 81.0% 2026-11-09
4 GPT-6 Astra Max OpenAI 80.6% 2026-11-09
5 GPT-6.1 Sol Max OpenAI 77.3% 2026-11-09
6 Muse Wonder Max Meta 72.4% 2026-11-09
7 Gemini 4 Argon High Google 70.8% 2026-11-09
8 DeepSeek V4.1 Flash 1017 Max DeepSeek 67.5% 2026-11-09
9 Muse Wonder Full Max Meta 62.8% 2026-11-09
10 DeepSeek V4.1 Flash Max DeepSeek 58.9% 2026-11-09
11 GLM-5.3 Max Zhipu AI 55.1% 2026-11-09
12 Grok 4.7 Extra High xAI 53.2% 2026-11-09
13 Claude Haiku 5.5 Max Anthropic 52.4% 2026-11-09
14 DeepSeek V4 Pro 0813 Max DeepSeek 52.0% 2026-11-09
15 GLM-5.3-Flash Max Zhipu AI 51.6% 2026-11-09
16 MiMo-V2.6-Pro Xiaomi 48.8% 2026-11-09
17 Step 5 Preview StepFun 47.0% 2026-11-09
18 GPT-6 Luna Max OpenAI 46.7% 2026-11-09
19 Hy4 preview Tencent 45.4% 2026-11-09
20 MiMo-V2.6-Flash Xiaomi 39.6% 2026-11-09
21 Mercury 3 High Inception 39.3% 2026-11-09
22 Gemini 3.8 Flash High Google 38.6% 2026-11-09
23 Kimi K3 Max Moonshot AI 28.6% 2026-11-09
24 Muse Spark Open Extra High Meta 26.9% 2026-11-09
25 Beam 1.0 Reflection 22.8% 2026-11-09
26 MiniMax-M3 MiniMax 20.9% 2026-11-09
27 Nemotron 3.5 Super VL 120B A12B High NVIDIA 16.4% 2026-11-09
28 Inkling Small Thinking Machines 10.4% 2026-11-09
29 Inkling Thinking Machines 8.8% 2026-11-09