# Vals-hosted Terminal-Bench 2.0 mirror (Vals Terminal-Bench 2.0 mirror)

> Vals AI hosted Terminal-Bench 2.0 view with easy, medium, and hard task splits.

Canonical page: https://benchlm.ai/benchmarks/valsterminalbench2

- Category: [Agentic](/agentic)
- Last updated: June 4, 2026

## About Vals Terminal-Bench 2.0 mirror

- Year: 2026
- Tasks: Terminal task difficulty splits
- Format: Accuracy score
- Difficulty: Terminal-based agent execution
- Paper: [Vals Terminal-Bench 2.0](https://www.vals.ai/benchmarks/terminal-bench-2)

BenchLM mirrors this Vals-hosted Terminal-Bench view as display-only secondary context.

Vals Terminal-Bench 2.0 mirror is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (67 models)

| Rank | Model | Configuration | Creator | Score |
|------|-------|---------------|---------|-------|
| 1 | [GPT-5.5](/models/gpt-5-5) | high reasoning | OpenAI | 73.20% |
| 2 | [Claude Opus 4.8](/models/claude-opus-4-8) | — | Anthropic | 70.04% |
| 3 | [Claude Opus 4.7](/models/claude-opus-4-7) | — | Anthropic | 68.54% |
| 4 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | high reasoning | Google | 67.42% |
| 5 | [Gemini 3.1 Pro Preview](https://www.vals.ai/models/google_gemini-3.1-pro-preview) | high reasoning | Google | 67.42% |
| 6 | [GPT-5.3 Codex](/models/gpt-5-3-codex) | high reasoning | OpenAI | 64.05% |
| 7 | [Muse Spark](/models/muse-spark) | — | Meta | 59.55% |
| 8 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | — | Anthropic | 59.55% |
| 9 | [Qwen3.7 Max](/models/qwen3-7-max) | — | Alibaba | 59.18% |
| 10 | [Claude Opus 4.5](/models/claude-opus-4-5) | — | Anthropic | 58.43% |
| 11 | [Claude Opus 4.6 (Adaptive)](/models/claude-opus-4-6-thinking) | — | Anthropic | 58.43% |
| 12 | [GPT-5.4](/models/gpt-5-4) | high reasoning | OpenAI | 58.43% |
| 13 | [Kimi K2.6](/models/kimi-2-6) | — | Moonshot AI | 57.30% |
| 14 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | — | DeepSeek | 56.18% |
| 15 | [Gemini 3 Pro Preview](https://www.vals.ai/models/google_gemini-3-pro-preview) | high reasoning | Google | 55.06% |
| 16 | [Claude Opus 4.5 Thinking](/models/claude-opus-4-5-thinking) | — | Anthropic | 53.93% |
| 17 | [GLM-5.1](/models/glm-5-1) | — | Z.AI | 53.93% |
| 18 | [Gemini 3 Flash Preview](https://www.vals.ai/models/google_gemini-3-flash-preview) | high reasoning | Google | 51.69% |
| 19 | [GPT-5.2](/models/gpt-5-2) | high reasoning | OpenAI | 51.69% |
| 20 | [Qwen 3.6 Max (preview)](/models/qwen3-6-max-preview) | — | Alibaba | 51.69% |
| 21 | [GLM 5 Thinking](https://www.vals.ai/models/zai_glm-5-thinking) | — | Zhipu AI | 49.44% |
| 22 | [MiniMax M2.7](/models/minimax-m2-7) | — | MiniMax | 47.19% |
| 23 | [MiniMax M3](/models/minimax-m3) | — | MiniMax | 46.07% |
| 24 | [Qwen3.6 Plus](/models/qwen3-6-plus) | — | Alibaba | 44.94% |
| 25 | [GPT-5.1](/models/gpt-5-1) | high reasoning | OpenAI | 44.94% |
| 26 | [Qwen3.6-27B](/models/qwen3-6-27b) | — | Alibaba | 44.94% |
| 27 | [GPT-5.4 mini](/models/gpt-5-4-mini) | high reasoning | OpenAI | 44.94% |
| 28 | [Grok 4.3](/models/grok-4-3) | — | xAI | 43.45% |
| 29 | [Claude Sonnet 4.5 Thinking](/models/claude-sonnet-4-5-thinking) | — | Anthropic | 41.57% |
| 30 | [MiniMax M2.5](/models/minimax-m2-5) | — | MiniMax | 41.57% |
| 31 | [Qwen3.5 Plus Thinking](https://www.vals.ai/models/alibaba_qwen3.5-plus-thinking) | — | Alibaba | 41.57% |
| 32 | [Grok 4.20 0309 Reasoning](https://www.vals.ai/models/grok_grok-4.20-0309-reasoning) | — | xAI | 40.45% |
| 33 | [Kimi K2.5 Thinking](https://www.vals.ai/models/kimi_kimi-k2.5-thinking) | — | Moonshot AI | 40.45% |
| 34 | [GPT-5.4 nano](/models/gpt-5-4-nano) | high reasoning | OpenAI | 39.89% |
| 35 | [Gemma 4 31b It](https://www.vals.ai/models/google_gemma-4-31b-it) | high reasoning | Google | 39.33% |
| 36 | [Claude Haiku 4.5 Thinking](/models/claude-haiku-4-5-thinking) | — | Anthropic | 38.20% |
| 37 | [GLM-4.7](/models/glm-4-7) | — | Z.AI | 38.20% |
| 38 | [MiniMax M2.1](https://www.vals.ai/models/minimax_MiniMax-M2.1) | — | MiniMax | 37.08% |
| 39 | [GPT-5](https://www.vals.ai/models/openai_gpt-5-2025-08-07) | high reasoning | OpenAI | 37.08% |
| 40 | [Kimi K2 Thinking](https://www.vals.ai/models/kimi_kimi-k2-thinking) | — | Moonshot AI | 37.08% |
| 41 | [DeepSeek V3p2 Thinking](https://www.vals.ai/models/fireworks_deepseek-v3p2-thinking) | high reasoning | Fireworks AI | 35.95% |
| 42 | [DeepSeek V3p2](https://www.vals.ai/models/fireworks_deepseek-v3p2) | none reasoning | Fireworks AI | 34.83% |
| 43 | [Laguna M.1](/models/laguna-m-1) | — | Poolside | 31.46% |
| 44 | [Gemini 2.5 Pro](/models/gemini-2-5-pro) | — | Google | 30.34% |
| 45 | [Mistral Medium 3.5](https://www.vals.ai/models/mistralai_mistral-medium-3.5) | high reasoning | Mistral AI | 30.34% |
| 46 | [Grok 4 Fast (Reasoning)](/models/grok-4-fast-reasoning) | — | xAI | 29.21% |
| 47 | [Laguna XS.2](/models/laguna-xs-2) | — | Poolside | 28.09% |
| 48 | [GLM-4.6](/models/glm-4-6) | — | Z.AI | 28.09% |
| 49 | [Grok 4 0709](https://www.vals.ai/models/grok_grok-4-0709) | — | xAI | 28.09% |
| 50 | [GPT-5 mini](/models/gpt-5-mini) | high reasoning | OpenAI | 26.97% |
| 51 | [Moonshotai Kimi K2 Instruct](https://www.vals.ai/models/together_moonshotai/Kimi-K2-Instruct) | — | Together AI | 25.84% |
| 52 | [Grok 4.1 Fast (Reasoning)](/models/grok-4-1-fast-reasoning) | — | xAI | 24.72% |
| 53 | [Gemini 3.1 Flash Lite Preview](https://www.vals.ai/models/google_gemini-3.1-flash-lite-preview) | high reasoning | Google | 24.72% |
| 54 | [Qwen3.5 Flash](/models/qwen3-5-flash) | — | Alibaba | 24.72% |
| 55 | [Qwen3 Max](/models/qwen3-max) | — | Alibaba | 24.72% |
| 56 | [DeepSeek V3p1](https://www.vals.ai/models/fireworks_deepseek-v3p1) | — | Fireworks AI | 22.47% |
| 57 | [Gemini 2.5 Flash Preview 09 2025 Thinking](https://www.vals.ai/models/google_gemini-2.5-flash-preview-09-2025-thinking) | — | Google | 21.35% |
| 58 | [Qwen3 Max](/models/qwen3-max) | — | Alibaba | 20.23% |
| 59 | [GPT-OSS 120B](/models/gpt-oss-120b) | — | OpenAI | 19.10% |
| 60 | [Grok 4.1 Fast Non Reasoning](https://www.vals.ai/models/grok_grok-4-1-fast-non-reasoning) | — | xAI | 17.98% |
| 61 | [Trinity-Large-Thinking](/models/trinity-large-thinking) | — | Arcee AI | 17.98% |
| 62 | [Mistral Small 2603](https://www.vals.ai/models/mistralai_mistral-small-2603) | high reasoning | Mistral AI | 16.85% |
| 63 | [GPT-4.1](/models/gpt-4-1) | high reasoning | OpenAI | 14.61% |
| 64 | [Magistral Medium 2509](https://www.vals.ai/models/mistralai_magistral-medium-2509) | — | Mistral AI | 13.48% |
| 65 | [Mistral Large 2512](https://www.vals.ai/models/mistralai_mistral-large-2512) | — | Mistral AI | 8.99% |
| 66 | [Command A 03 2025](https://www.vals.ai/models/cohere_command-a-03-2025) | — | Cohere | 2.25% |
| 67 | [Llama4 Maverick Instruct Basic](https://www.vals.ai/models/fireworks_llama4-maverick-instruct-basic) | — | Fireworks AI | 2.25% |

## FAQ

### What does Vals Terminal-Bench 2.0 mirror measure?

Vals AI hosted Terminal-Bench 2.0 view with easy, medium, and hard task splits.

### Which model leads the published Vals Terminal-Bench 2.0 mirror snapshot?

GPT-5.5 currently leads the published Vals Terminal-Bench 2.0 mirror snapshot with a score of 73.20%.

### How many models are evaluated on Vals Terminal-Bench 2.0 mirror?

The June 4, 2026 contains 67 AI models.

### Does Vals Terminal-Bench 2.0 mirror affect BenchLM's overall score?

Not directly. Vals Terminal-Bench 2.0 mirror is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.