TestsBenchmarks · Tool use
MCP Atlas
Share of realistic multi-step tasks solved by calling real MCP tool servers.
Measured by independent testersIndependent Reported by the makerVendor-reported
| Thinking levelSetting | |||||||
|---|---|---|---|---|---|---|---|
| 1 | Muse Spark 1.2 | Meta (Meta Superintelligence Labs, Muse) | 90.3% | Extra high | Maker's own figureVendor-reportedMeta ↗ | — | 5 Aug 2026 |
| 2 | Muse Spark 1.1 | Meta (Meta Superintelligence Labs, Muse) | 88.1% | DefaultMuse Spark 1.1 | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 3 | Claude Fable 5.1 | Anthropic | 87.2% | DefaultFable 5.1 | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 4 | Claude Opus 5 | Anthropic | 85.8% | Extra highclaude-opus-5 (xhigh) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 5 | Qwen3.8 2.4T A95B (open weights) | Qwen (Alibaba) | 84.5% | Extra highQwen3.8-2.4T-A95B (xHigh) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 6 | GLM-5.3 | Z.ai (Zhipu) | 84.2% | DefaultGLM 5.3 | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 7 | Gemini 3.5 Flash | Google (Gemini / DeepMind) | 83.6% | HighGemini 3.5 Flash (high) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 8 | Claude Fable 5 | Anthropic | 83.3% | DefaultClaude Fable 5 | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 9 | Kimi K3 | Moonshot AI (Kimi) | 82.3% | Maxkimi-k3 (max) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 10 | Claude Opus 4.8 | Anthropic | 82.2% | Maxclaude-opus-4-8 (max) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 11 | Muse Spark | Meta (Meta Superintelligence Labs, Muse) | 82.2% | DefaultMuse Spark | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 12 | GPT-5.6 Sol | OpenAI | 81.8% | Defaultgpt-5.6 (sol) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 13 | Inkling Small | Thinking Machines Lab | 79.2% | DefaultInkling-small | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 14 | Claude Opus 4.7 | Anthropic | 79.1% | Maxclaude-opus-4-7 (max) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 15 | Gemini 3.1 Pro (Preview) | Google (Gemini / DeepMind) | 78.2% | Highgemini-3.1-pro-preview (high) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 16 | GLM-5.2 | Z.ai (Zhipu) | 77.8% | Defaultglm-5p2 | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 17 | Claude Opus 4.6 | Anthropic | 76.8% | Maxclaude-opus-4-6 (max) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 18 | Qwen3.7 Max | Qwen (Alibaba) | 76.4% | Defaultthinking (vendor recommends 'Reasoning effort is set to xhigh' system prompt for reasoning tasks) | Maker's own figureVendor-reportedQwen (Alibaba) ↗ | — | 16 May 2026 |
| 19 | Inkling | Thinking Machines Lab | 76.0% | Extra highInkling (xHigh) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 20 | Kimi K2.7 Code | Moonshot AI (Kimi) | 76.0% | Defaultthinking | Maker's own figureVendor-reportedMoonshot AI ↗ | Kimi Code CLI | 12 Jun 2026 |
| 21 | GLM-5.1 | Z.ai (Zhipu) | 75.6% | Defaultglm-5p1 | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 22 | Muse Glimmer 30B | Meta (Meta Superintelligence Labs, Muse) | 75.5% | HighHigh reasoning | Maker's own figureVendor-reportedMeta ↗ | — | 10 Aug 2026 |
| 23 | GPT-5.5 | OpenAI | 75.3% | Extra highgpt-5.5 (xhigh) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 24 | DeepSeek V4 Pro (Preview, 0423) | DeepSeek | 74.2% | HighThink High | Maker's own figureVendor-reportedDeepSeek ↗ | — | 24 Apr 2026 |
| 25 | MiniMax M3 | MiniMax | 74.2% | Defaultsetting not stated | Maker's own figureVendor-reportedMiniMax ↗ | — | 1 Jun 2026 |
| 26 | Qwen3.6 Plus | Qwen (Alibaba) | 74.1% | Defaultthinking | Maker's own figureVendor-reportedQwen (Alibaba) ↗ | — | 2 Apr 2026 |
| 27 | Qwen3.7 Plus | Qwen (Alibaba) | 73.2% | Defaultthinking | Maker's own figureVendor-reportedQwen (Alibaba) ↗ | — | 21 May 2026 |
| 28 | GPT-5.4 | OpenAI | 70.6% | Extra highgpt-5.4 (xhigh) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 29 | Gemini 3 Pro | Google (Gemini / DeepMind) | 70.3% | Defaultgemini-3-pro-preview | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 30 | Claude Opus 4.5 | Anthropic | 69.8% | Highclaude-opus-4-5 (high) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 31 | Claude Sonnet 4.6 | Anthropic | 69.5% | Defaultclaude-sonnet-4-6 | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 32 | Kimi K2.6 | Moonshot AI (Kimi) | 69.4% | Defaultthinking | Maker's own figureVendor-reportedMoonshot AI ↗ | Kimi Code CLI | 12 Jun 2026 |
| 33 | DeepSeek V4 Flash (Preview, 0423) | DeepSeek | 69.0% | MaxThink Max | Maker's own figureVendor-reportedDeepSeek ↗ | — | 24 Apr 2026 |
| 34 | GPT-5.2 | OpenAI | 67.6% | Extra highgpt-5.2 (xhigh) | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 35 | Kimi K2.5 | Moonshot AI (Kimi) | 64.4% | Defaultkimi-k2p5 | Independent testIndependentScale AI SEAL ↗ | — | 1 Oct 2026 |
| 36 | MiniMax M2.7 | MiniMax | 49.4% | Defaultsetting not stated | Maker's own figureVendor-reportedMiniMax ↗ | — | 1 Jun 2026 |