# Vals-hosted SWE-bench mirror (Vals SWE-bench mirror)

> Vals AI hosted SWE-bench view for solving production software engineering tasks.

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

- Category: [external](/external)
- Last updated: September 1, 2026

## About Vals SWE-bench mirror

- Year: 2026
- Tasks: Software engineering issue-resolution tasks
- Format: Accuracy score
- Difficulty: Production software engineering
- Paper: [Vals SWE-bench](https://www.vals.ai/benchmarks/swebench)

BenchLM keeps this separate from its canonical SWE-bench Verified page so Vals-hosted results remain secondary context rather than source-of-record data.

Vals SWE-bench mirror is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (88 models)

| Rank | Model | Configuration | Creator | Score |
|------|-------|---------------|---------|-------|
| 1 | [Claude Opus 5](/models/claude-opus-5) | — | Anthropic | 97.00% |
| 2 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | max reasoning | DeepSeek | 96.40% |
| 3 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | max reasoning | OpenAI | 96.20% |
| 4 | [Grok 4.6](/models/grok-4-6) | high reasoning | xAI | 95.60% |
| 5 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | max reasoning | OpenAI | 95.40% |
| 6 | [GLM-5.3](/models/glm-5-3) | max reasoning | Z.AI | 95.40% |
| 7 | [Claude Fable 5](/models/claude-fable) | — | Anthropic | 95.00% |
| 8 | [Kimi K3](/models/kimi-k3) | — | Moonshot AI | 93.40% |
| 9 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | max reasoning | OpenAI | 93.00% |
| 10 | [GLM-5.3-Flash](/models/glm-5-3-flash) | max reasoning | Z.AI | 92.00% |
| 11 | [DeepSeek V4 Flash 0731](/models/deepseek-v4-flash-0731) | high reasoning | DeepSeek | 88.80% |
| 12 | [Claude Opus 4.8](/models/claude-opus-4-8) | — | Anthropic | 88.60% |
| 13 | [Grok 4.5](/models/grok-4-5) | high reasoning | xAI | 86.60% |
| 14 | [Muse Spark 1.2](/models/muse-spark-1-2) | xhigh reasoning | Meta | 86.60% |
| 15 | [Qwen3.8-27B](/models/qwen3-8-27b) | xhigh reasoning | Alibaba | 86.00% |
| 16 | [Claude Opus 4.8 Claude Code](https://www.vals.ai/models/anthropic_claude-opus-4-8-claude-code) | Claude Code · Claude Code | Anthropic | 85.80% |
| 17 | [Qwen3.8 Max](/models/qwen3-8-max) | — | Alibaba | 85.60% |
| 18 | [GLM-5.2](/models/glm-5-2) | max reasoning | Z.AI | 82.80% |
| 19 | [GPT-5.5](/models/gpt-5-5) | xhigh reasoning | OpenAI | 82.60% |
| 20 | [Inkling-Small](/models/inkling-small) | 0.99 reasoning | Thinking Machines Lab | 82.20% |
| 21 | [Muse Spark 1.1](/models/muse-spark-1-1) | xhigh reasoning | Meta | 82.00% |
| 22 | [Claude Opus 4.7](/models/claude-opus-4-7) | — | Anthropic | 82.00% |
| 23 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | high reasoning | Google | 80.80% |
| 24 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | high reasoning | Google | 80.00% |
| 25 | [Composer 2.5](/models/composer-2-5) | Cursor CLI · Cursor CLI | Cursor | 79.60% |
| 26 | [Gemini 3.6 Flash](/models/gemini-3-6-flash) | high reasoning | Google | 79.60% |
| 27 | [Claude Sonnet 5](/models/claude-sonnet-5) | — | Anthropic | 79.60% |
| 28 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | high reasoning | Google | 78.80% |
| 29 | [Gemini 3.1 Pro Preview](https://www.vals.ai/models/google_gemini-3.1-pro-preview) | high reasoning | Google | 78.80% |
| 30 | [GPT-5.4](/models/gpt-5-4) | xhigh reasoning | OpenAI | 78.20% |
| 31 | [Claude Opus 4.6 (Adaptive)](/models/claude-opus-4-6-thinking) | — | Anthropic | 78.20% |
| 32 | [Kimi K2.7 Code](/models/kimi-k2-7-code) | — | Moonshot AI | 78.20% |
| 33 | [GPT-5.3 Codex](/models/gpt-5-3-codex) | xhigh reasoning | OpenAI | 78.00% |
| 34 | [Inkling](/models/inkling) | 0.99 reasoning | Thinking Machines Lab | 77.60% |
| 35 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | — | Anthropic | 77.40% |
| 36 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | max reasoning | DeepSeek | 77.40% |
| 37 | [GPT-5.5 Codex](https://www.vals.ai/models/openai_gpt-5.5-codex) | Codex · Codex | OpenAI | 76.40% |
| 38 | [Claude Opus 4.5 Thinking](/models/claude-opus-4-5-thinking) | — | Anthropic | 76.40% |
| 39 | [Gemini 3 Pro Preview](https://www.vals.ai/models/google_gemini-3-pro-preview) | high reasoning | Google | 76.40% |
| 40 | [GLM-5.1](/models/glm-5-1) | — | Z.AI | 76.40% |
| 41 | [GPT-5.5 Factory](https://www.vals.ai/models/openai_gpt-5.5-factory) | Factory · Factory | OpenAI | 76.20% |
| 42 | [Kimi K2.6](/models/kimi-2-6) | — | Moonshot AI | 76.20% |
| 43 | [GPT-5.2](/models/gpt-5-2) | xhigh reasoning | OpenAI | 75.80% |
| 44 | [Gemini 3 Flash Preview](https://www.vals.ai/models/google_gemini-3-flash-preview) | high reasoning | Google | 75.00% |
| 45 | [Gemini 3.5 Flash-Lite](/models/gemini-3-5-flash-lite) | high reasoning | Google | 75.00% |
| 46 | [MiniMax M3](/models/minimax-m3) | — | MiniMax | 75.00% |
| 47 | [MiniMax M2.1](https://www.vals.ai/models/minimax_MiniMax-M2.1) | — | MiniMax | 74.80% |
| 48 | [Muse Spark](/models/muse-spark) | — | Meta | 74.40% |
| 49 | [MiniMax M2.5](/models/minimax-m2-5) | — | MiniMax | 74.20% |
| 50 | [MiMo-V2.5-Pro](/models/mimo-v2-5-pro) | — | Xiaomi | 74.00% |
| 51 | [MiniMax M2.7](/models/minimax-m2-7) | — | MiniMax | 73.80% |
| 52 | [Qwen3.6 Plus](/models/qwen3-6-plus) | — | Alibaba | 73.40% |
| 53 | [GPT-5.4 mini](/models/gpt-5-4-mini) | xhigh reasoning | OpenAI | 73.00% |
| 54 | [Qwen 3.6 Max (preview)](/models/qwen3-6-max-preview) | — | Alibaba | 72.80% |
| 55 | [GPT-5.2-Codex](/models/gpt-5-2-codex) | high reasoning | OpenAI | 72.40% |
| 56 | [Grok 4.20 0309 Reasoning](https://www.vals.ai/models/grok_grok-4.20-0309-reasoning) | — | xAI | 72.20% |
| 57 | [Grok 4.3](/models/grok-4-3) | high reasoning | xAI | 71.40% |
| 58 | [GLM 5 Thinking](https://www.vals.ai/models/zai_glm-5-thinking) | — | Zhipu AI | 71.40% |
| 59 | [Qwen3.5 Plus Thinking](https://www.vals.ai/models/alibaba_qwen3.5-plus-thinking) | — | Alibaba | 71.20% |
| 60 | [MiMo-V2.5](/models/mimo-v2-5) | — | Xiaomi | 71.00% |
| 61 | [Kimi K2.5 Thinking](https://www.vals.ai/models/kimi_kimi-k2.5-thinking) | — | Moonshot AI | 70.00% |
| 62 | [Claude Sonnet 4.5 Thinking](/models/claude-sonnet-4-5-thinking) | — | Anthropic | 70.00% |
| 63 | [Qwen3.6-27B](/models/qwen3-6-27b) | — | Alibaba | 70.00% |
| 64 | [GPT-5.4 nano](/models/gpt-5-4-nano) | high reasoning | OpenAI | 69.80% |
| 65 | [GPT-5.1](/models/gpt-5-1) | high reasoning | OpenAI | 69.80% |
| 66 | [GLM-4.7](/models/glm-4-7) | — | Z.AI | 69.40% |
| 67 | [GPT-5](https://www.vals.ai/models/openai_gpt-5-2025-08-07) | high reasoning | OpenAI | 69.00% |
| 68 | [Nemotron 3 Ultra 550b A55b](https://www.vals.ai/models/nvidia_nemotron-3-ultra-550b-a55b) | — | Nvidia | 69.00% |
| 69 | [Qwen3.7 Max](/models/qwen3-7-max) | — | Alibaba | 68.80% |
| 70 | [DeepSeek V3p2 Thinking](https://www.vals.ai/models/fireworks_deepseek-v3p2-thinking) | high reasoning | Fireworks AI | 67.60% |
| 71 | [Claude Haiku 4.5 Thinking](/models/claude-haiku-4-5-thinking) | — | Anthropic | 66.60% |
| 72 | [Mistral Medium 3.5](https://www.vals.ai/models/mistralai_mistral-medium-3.5) | high reasoning | Mistral AI | 66.40% |
| 73 | [Ling 3.0 Flash 2607](https://www.vals.ai/models/ant_ling-3.0-flash-2607) | — | Ant | 65.20% |
| 74 | [Qwen3.5 Flash](/models/qwen3-5-flash) | — | Alibaba | 64.40% |
| 75 | [Gemini 3.1 Flash Lite Preview](https://www.vals.ai/models/google_gemini-3.1-flash-lite-preview) | high reasoning | Google | 62.80% |
| 76 | [Devstral 2512](https://www.vals.ai/models/mistralai_devstral-2512) | — | Mistral AI | 62.80% |
| 77 | [GPT-5 mini](/models/gpt-5-mini) | high reasoning | OpenAI | 60.80% |
| 78 | [Kimi K2 Thinking](https://www.vals.ai/models/kimi_kimi-k2-thinking) | — | Moonshot AI | 60.20% |
| 79 | [Grok 4 0709](https://www.vals.ai/models/grok_grok-4-0709) | — | xAI | 57.80% |
| 80 | [Laguna M.1](/models/laguna-m-1) | — | Poolside | 57.60% |
| 81 | [Laguna XS.2](/models/laguna-xs-2) | — | Poolside | 55.20% |
| 82 | [Gemini 2.5 Pro](/models/gemini-2-5-pro) | — | Google | 54.40% |
| 83 | [Nemotron Lightning 3p5 30b A3b](https://www.vals.ai/models/fireworks_nemotron-lightning-3p5-30b-a3b) | — | Fireworks AI | 52.80% |
| 84 | [Grok 4 Fast (Reasoning)](/models/grok-4-fast-reasoning) | — | xAI | 45.40% |
| 85 | [Grok 4.1 Fast (Reasoning)](/models/grok-4-1-fast-reasoning) | — | xAI | 41.40% |
| 86 | [Mistral Large 2512](https://www.vals.ai/models/mistralai_mistral-large-2512) | — | Mistral AI | 41.40% |
| 87 | [GPT-OSS 120B](/models/gpt-oss-120b) | — | OpenAI | 33.60% |
| 88 | [Command A 03 2025](https://www.vals.ai/models/cohere_command-a-03-2025) | — | Cohere | 7.80% |

## FAQ

### What does Vals SWE-bench mirror measure?

Vals AI hosted SWE-bench view for solving production software engineering tasks.

### Which model leads the published Vals SWE-bench mirror snapshot?

Claude Opus 5 currently leads the published Vals SWE-bench mirror snapshot with a score of 97.00%.

### How many models are evaluated on Vals SWE-bench mirror?

The September 1, 2026 contains 88 AI models.

### Does Vals SWE-bench mirror affect BenchLM's overall score?

Not directly. Vals SWE-bench mirror is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.