# Vals Code Migration (Code Migration)

> Can language models reimplement real-world programs in another language?

Canonical page: https://benchlm.ai/benchmarks/codemigration

- Category: [Coding](/coding)
- Last updated: September 27, 2026

## About Code Migration

- Year: 2026
- Tasks: Real-world program reimplementation in another language
- Format: Accuracy score
- Difficulty: Production code migration
- Paper: [Code Migration](https://www.vals.ai/benchmarks/code-migration)

BenchLM mirrors the public Vals AI Code Migration leaderboard as display-only external evidence. The captured snapshot preserves overall scores, task-level scores where Vals publishes them, uncertainty, latency, and cost-per-test metadata. It is excluded from BenchLM weighted rankings.

Code Migration is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (69 models)

| Rank | Model | Configuration | Creator | Score |
|------|-------|---------------|---------|-------|
| 1 | [Claude Sonnet 5.5](/models/claude-sonnet-5-5) | — | Anthropic | 69.83% |
| 2 | [GPT-6 Astra](/models/gpt-6-astra) | max reasoning | OpenAI | 67.74% |
| 3 | [Claude Opus 5.5](/models/claude-opus-5-5) | — | Anthropic | 66.65% |
| 4 | [Claude Opus 5](/models/claude-opus-5) | — | Anthropic | 57.47% |
| 5 | [GPT-6 Sol](/models/gpt-6-sol) | max reasoning | OpenAI | 57.20% |
| 6 | [Claude Fable 5](/models/claude-fable) | — | Anthropic | 55.06% |
| 7 | [Claude Fable 5.1](/models/claude-fable-5-1) | — | Anthropic | 54.61% |
| 8 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | max reasoning | OpenAI | 52.92% |
| 9 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | max reasoning | OpenAI | 47.80% |
| 10 | [Hy4 preview](/models/hy4-preview) | — | Tencent | 47.43% |
| 11 | [Muse Spark 1.3 Max](https://www.vals.ai/models/meta_muse_spark_1_3_max) | max reasoning | Meta | 47.41% |
| 12 | [Claude Opus 4.8](/models/claude-opus-4-8) | — | Anthropic | 47.25% |
| 13 | [DeepSeek V4.1 Flash](/models/deepseek-v4-1-flash) | high reasoning | DeepSeek | 45.62% |
| 14 | [GPT-5.5](/models/gpt-5-5) | xhigh reasoning | OpenAI | 45.16% |
| 15 | [Grok 4.7](/models/grok-4-7) | xhigh reasoning | xAI | 44.82% |
| 16 | [Grok 4.6](/models/grok-4-6) | high reasoning | xAI | 44.57% |
| 17 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | max reasoning | OpenAI | 44.55% |
| 18 | [Claude Sonnet 5](/models/claude-sonnet-5) | — | Anthropic | 44.39% |
| 19 | [GLM-5.3](/models/glm-5-3) | max reasoning | Z.AI | 44.22% |
| 20 | [Claude Opus 4.7](/models/claude-opus-4-7) | — | Anthropic | 43.88% |
| 21 | [MiMo-V2.6-Pro](/models/mimo-v2-6-pro) | — | Xiaomi | 43.01% |
| 22 | [GPT-6 Luna](/models/gpt-6-luna) | max reasoning | OpenAI | 42.55% |
| 23 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | max reasoning | DeepSeek | 41.54% |
| 24 | [MiMo-V2.6-Flash](/models/mimo-v2-6-flash) | — | Xiaomi | 40.93% |
| 25 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | — | Anthropic | 39.89% |
| 26 | [DeepSeek V4 Flash 0731](/models/deepseek-v4-flash-0731) | high reasoning | DeepSeek | 38.63% |
| 27 | [GLM-5.2](/models/glm-5-2) | max reasoning | Z.AI | 37.87% |
| 28 | [Grok 4.5](/models/grok-4-5) | high reasoning | xAI | 36.60% |
| 29 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | high reasoning | Google | 36.55% |
| 30 | [GPT-5.4](/models/gpt-5-4) | xhigh reasoning | OpenAI | 34.98% |
| 31 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | high reasoning | Google | 34.80% |
| 32 | [Muse Spark 1.1](/models/muse-spark-1-1) | xhigh reasoning | Meta | 31.11% |
| 33 | [Gemini 3.6 Flash](/models/gemini-3-6-flash) | high reasoning | Google | 30.93% |
| 34 | [Muse Spark 1.2](/models/muse-spark-1-2) | xhigh reasoning | Meta | 29.95% |
| 35 | [Kimi K2.6](/models/kimi-2-6) | — | Moonshot AI | 27.77% |
| 36 | [Muse Spark 1.3](/models/muse-spark-1-3) | xhigh reasoning | Meta | 27.58% |
| 37 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | high reasoning | Google | 26.75% |
| 38 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | max reasoning | DeepSeek | 26.20% |
| 39 | [GLM-5.1](/models/glm-5-1) | — | Z.AI | 25.77% |
| 40 | [Kimi K2.7 Code](/models/kimi-k2-7-code) | — | Moonshot AI | 25.39% |
| 41 | [Qwen3.8 Max](/models/qwen3-8-max) | — | Alibaba | 23.96% |
| 42 | [MiMo-V2.5-Pro](/models/mimo-v2-5-pro) | — | Xiaomi | 21.56% |
| 43 | [GLM-5.3-Flash](/models/glm-5-3-flash) | max reasoning | Z.AI | 20.52% |
| 44 | [MiniMax M3](/models/minimax-m3) | — | MiniMax | 19.93% |
| 45 | [Gemini 3.1 Pro Preview](https://www.vals.ai/models/google_gemini-3.1-pro-preview) | high reasoning | Google | 17.31% |
| 46 | [Kimi K3](/models/kimi-k3) | max reasoning | Moonshot AI | 16.10% |
| 47 | [GPT-5.4 nano](/models/gpt-5-4-nano) | high reasoning | OpenAI | 14.47% |
| 48 | [MiMo-V2.5](/models/mimo-v2-5) | — | Xiaomi | 14.23% |
| 49 | [Qwen3.8-27B](/models/qwen3-8-27b) | xhigh reasoning | Alibaba | 14.16% |
| 50 | [Inkling-Small](/models/inkling-small) | 0.99 reasoning | Thinking Machines Lab | 13.68% |
| 51 | [Qwen3.7 Max](/models/qwen3-7-max) | — | Alibaba | 13.07% |
| 52 | [GPT-5.4 mini](/models/gpt-5-4-mini) | xhigh reasoning | OpenAI | 12.94% |
| 53 | [Qwen3.7 Plus](/models/qwen3-7-plus) | — | Alibaba | 12.86% |
| 54 | [Inkling](/models/inkling) | 0.99 reasoning | Thinking Machines Lab | 11.79% |
| 55 | [Qwen3.6 Plus](/models/qwen3-6-plus) | — | Alibaba | 11.10% |
| 56 | [MiniMax M2.7](/models/minimax-m2-7) | — | MiniMax | 8.69% |
| 57 | [Claude Haiku 4.5 Thinking](/models/claude-haiku-4-5-thinking) | — | Anthropic | 7.55% |
| 58 | [Kimi K2.5 Thinking](https://www.vals.ai/models/kimi_kimi-k2.5-thinking) | — | Moonshot AI | 6.96% |
| 59 | [Grok 4.3](/models/grok-4-3) | high reasoning | xAI | 6.79% |
| 60 | [Gemini 3 Flash Preview](https://www.vals.ai/models/google_gemini-3-flash-preview) | high reasoning | Google | 6.39% |
| 61 | [Gemini 3.5 Flash-Lite](/models/gemini-3-5-flash-lite) | high reasoning | Google | 6.15% |
| 62 | [Mistral Medium 3.5](https://www.vals.ai/models/mistralai_mistral-medium-3.5) | high reasoning | Mistral AI | 5.13% |
| 63 | [Nemotron 3 Ultra 550b A55b](https://www.vals.ai/models/nvidia_nemotron-3-ultra-550b-a55b) | — | Nvidia | 4.91% |
| 64 | [Gemini 3.1 Flash Lite Preview](https://www.vals.ai/models/google_gemini-3.1-flash-lite-preview) | high reasoning | Google | 4.61% |
| 65 | [Mercury 2.5](/models/mercury-2-5) | high reasoning | Inception | 4.49% |
| 66 | [Nemotron Lightning 3p5 30b A3b](https://www.vals.ai/models/fireworks_nemotron-lightning-3p5-30b-a3b) | — | Fireworks AI | 2.95% |
| 67 | [Ling 3.0 Flash Af Rc3](https://www.vals.ai/models/ant_ling-3.0-flash-af-rc3) | — | Ant | 1.82% |
| 68 | [Grok 4.20 0309 Reasoning](https://www.vals.ai/models/grok_grok-4.20-0309-reasoning) | — | xAI | 0.31% |
| 69 | [Ling 3.0 Flash 2607](https://www.vals.ai/models/ant_ling-3.0-flash-2607) | — | Ant | 0.00% |

## FAQ

### What does Code Migration measure?

Can language models reimplement real-world programs in another language?

### Which model leads the published Code Migration snapshot?

Claude Sonnet 5.5 currently leads the published Code Migration snapshot with a score of 69.83%.

### How many models are evaluated on Code Migration?

The September 27, 2026 contains 69 AI models.

### Does Code Migration affect BenchLM's overall score?

Not directly. Code Migration is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.