TestsBenchmarks · Agentic
Terminal-Bench 4.0
Share of 66 hard, real terminal tasks (software, science, ML, ops, security, hardware) an agent completes end-to-end in a container.
Measured by independent testersIndependent Reported by the makerVendor-reported
| Thinking levelSetting | |||||||
|---|---|---|---|---|---|---|---|
| 1 | Claude Sonnet 5.5 | Anthropic | 66.2% | MaxSonnet 5.5 (max) | Independent testIndependentArtificial Analysis ↗ | Claude Code | 1 Oct 2026 |
| 2 | Claude Opus 5.5 | Anthropic | 63.1% | MaxOpus 5.5 (max) | Independent testIndependentArtificial Analysis ↗ | Claude Code | 1 Oct 2026 |
| 3 | GPT-6 Astra | OpenAI | 59.6% | Extra highGPT-6 Astra (Xhigh) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 4 | Claude Fable 5.1 | Anthropic | 57.9% | Max | Independent testIndependentTerminal-Bench ↗ | Claude Code | 1 Oct 2026 |
| 5 | Gemini 4 Argon | Google (Gemini / DeepMind) | 57.4% | Maxhighest thinking settings | Maker's own figureVendor-reportedGoogle ↗ | — | 30 Sep 2026 |
| 6 | GPT-6.1 Sol | OpenAI | 56.1% | MaxGPT-6.1 Sol (Max) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 7 | Claude Opus 5 | Anthropic | 54.5% | MaxOpus 5 (max) | Independent testIndependentArtificial Analysis ↗ | Claude Code | 1 Oct 2026 |
| 8 | Claude Fable 5 | Anthropic | 44.5% | Max | Independent testIndependentTerminal-Bench ↗ | Claude Code | 1 Oct 2026 |
| 9 | GPT-6 Sol | OpenAI | 43.9% | MaxGPT-6 Sol (Max) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 10 | GLM-5.3 | Z.ai (Zhipu) | 41.9% | MaxGLM-5.3 (Max) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 11 | GPT-5.6 Sol | OpenAI | 39.9% | MaxGPT-5.6 Sol (Max) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 12 | Qwen3.8 Max (0902) | Qwen (Alibaba) | 38.9% | DefaultQwen3.8 Max (0902) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 13 | Grok 4.7 | SpaceXAI (formerly xAI) | 37.6% | Extra high | Independent testIndependentTerminal-Bench ↗ | Grok Build | 1 Oct 2026 |
| 14 | GPT-5.6 Terra | OpenAI | 35.4% | MaxGPT-5.6 Terra (Max) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 15 | MiMo-V2.6-Pro | Xiaomi MiMo | 34.8% | DefaultMiMo-V2.6-Pro | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 16 | Step 5 Preview | StepFun | 33.3% | DefaultStep 5 Preview | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 17 | Muse Spark 1.3 | Meta (Meta Superintelligence Labs, Muse) | 33.3% | MaxMuse Spark 1.3 (Max) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 18 | GLM-5.3-Flash | Z.ai (Zhipu) | 32.8% | DefaultGLM 5.3 Flash | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 19 | DeepSeek V4.1 Flash | DeepSeek | 26.8% | MaxDeepSeek V4.1 Flash (Max) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 20 | Qwen3.8-Flash-Next | Qwen (Alibaba) | 25.3% | DefaultQwen3.8-Flash-Next | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 21 | Claude Opus 4.8 | Anthropic | 23.6% | Max | Independent testIndependentTerminal-Bench ↗ | Claude Code | 1 Oct 2026 |
| 22 | MiMo-V2.6-Flash | Xiaomi MiMo | 22.7% | DefaultMiMo-V2.6-Flash | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 23 | Grok 4.6 | SpaceXAI (formerly xAI) | 21.2% | HighGrok 4.6 (High) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 24 | Kimi K3 | Moonshot AI (Kimi) | 21.2% | DefaultKimi K3 | Independent testIndependentArtificial Analysis ↗ | Kimi Code CLI | 1 Oct 2026 |
| 25 | Ember-1 | Fireworks AI | 19.7% | Default | Independent testIndependentVals.ai ↗ | mini-SWE-agent | 29 Sep 2026 |
| 26 | Gemini 3.8 Flash | Google (Gemini / DeepMind) | 19.7% | MediumGemini 3.8 Flash (Medium) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 27 | DeepSeek V4 Flash (0731) | DeepSeek | 18.7% | Highreasoning_effort=high | Independent testIndependentVals.ai ↗ | mini-SWE-agent | 29 Sep 2026 |
| 28 | GPT-5.6 Luna | OpenAI | 17.3% | Max | Independent testIndependentTerminal-Bench ↗ | Codex | 1 Oct 2026 |
| 29 | GPT-6 Luna | OpenAI | 15.2% | MaxGPT-6 Luna (max) | Independent testIndependentArtificial Analysis ↗ | Codex | 1 Oct 2026 |
| 30 | GPT-5.5 | OpenAI | 14.6% | Extra highGPT-5.5 (Xhigh) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 31 | DeepSeek V4 Pro (0813) | DeepSeek | 14.1% | MaxDeepSeek V4 Pro 0813 (Max) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 32 | Claude Sonnet 5 | Anthropic | 14.1% | MaxClaude Sonnet 5 (Adaptive Reasoning, Max Effort) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 33 | Gemini 3.7 Flash | Google (Gemini / DeepMind) | 13.6% | HighGemini 3.7 Flash (High) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 34 | Grok 4.5 | SpaceXAI (formerly xAI) | 12.4% | High | Independent testIndependentTerminal-Bench ↗ | Grok Build | 1 Oct 2026 |
| 35 | DeepSeek V4 Flash Vision Exp | DeepSeek | 12.1% | MaxDeepSeek V4 Flash Vision (Max) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 36 | Qwen3.8 2.4T A95B (open weights) | Qwen (Alibaba) | 11.1% | DefaultQwen3.8 2.4T A95B | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 37 | Muse Spark 1.2 | Meta (Meta Superintelligence Labs, Muse) | 7.1% | Extra highMuse Spark 1.2 (Xhigh) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 38 | Gemini 3.6 Flash | Google (Gemini / DeepMind) | 7.1% | HighGemini 3.6 Flash (High) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 39 | Muse Spark 1.1 | Meta (Meta Superintelligence Labs, Muse) | 6.1% | Extra highMuse Spark 1.1 (Xhigh) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 40 | Qwen3.8 27B | Qwen (Alibaba) | 5.6% | Extra highQwen3.8 27B (Xhigh) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |
| 41 | GLM-5.2 | Z.ai (Zhipu) | 1.0% | MaxGLM-5.2 (Max) | Independent testIndependentArtificial Analysis ↗ | — | 1 Oct 2026 |