Benchmark
MCP Atlas Benchmark
What is MCP Atlas?
MCP Atlas is a benchmark for evaluating AI models on scaled tool use capabilities, measuring how well models can coordinate and utilize multiple tools across…
- Modality
- text
- Evaluates
- reasoning, agents, code, tool calling
- Openness
- unknown
- Importer
- LLM Stats
- Reported scores
- 33
Source provenance
- Original evidence https://api.zeroeval.com/leaderboard/benchmarks/mcp-atlas?top_n=500
MCP Atlas paper and code
MCP Atlas results and reported scores
LLM Stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| Claude Opus 4.5 | Anthropic | 0.623 | 2025-11-24 | evidence |
| Claude Opus 4.6 | Anthropic | 0.627 | 2026-02-05 | evidence |
| Claude Opus 4.7 | Anthropic | 0.773 | 2026-04-16 | evidence |
| Claude Opus 4.8 | Anthropic | 0.822 | 2026-05-28 | evidence |
| Claude Sonnet 4.6 | Anthropic | 0.613 | 2026-02-17 | evidence |
| DeepSeek-V4-Flash-0423 | DeepSeek | 0.674 | 2026-04-23 | evidence |
| DeepSeek-V4-Flash-Max | DeepSeek | 0.69 | 2026-04-23 | evidence |
| DeepSeek-V4-Pro-Max | DeepSeek | 0.736 | 2026-04-23 | evidence |
| Gemini 3 Flash | 0.574 | 2025-12-17 | evidence | |
| Gemini 3.1 Pro | 0.692 | 2026-02-19 | evidence | |
| Gemini 3.5 Flash | 0.836 | 2026-05-19 | evidence | |
| GLM-5 | Z.ai | 0.678 | 2026-02-11 | evidence |
| GLM-5.1 | Z.ai | 0.718 | 2026-04-07 | evidence |
| GLM-5.2 | Z.ai | 0.768 | 2026-06-16 | evidence |
| GPT-5.2 | OpenAI | 0.606 | 2025-12-11 | evidence |
| GPT-5.4 | OpenAI | 0.672 | 2026-03-05 | evidence |
| GPT-5.4 mini | OpenAI | 0.577 | 2026-03-17 | evidence |
| GPT-5.4 nano | OpenAI | 0.561 | 2026-03-17 | evidence |
| GPT-5.5 | OpenAI | 0.753 | 2026-04-23 | evidence |
| Hy3 | Tencent | 0.791 | 2026-07-06 | evidence |
| Inkling-Small | Thinking Machines Lab | 0.796 | 2026-07-30 | evidence |
| Kimi K2.7 Code | Moonshot AI | 0.76 | 2026-06-12 | evidence |
| Kimi K3 | Moonshot AI | 0.842 | 2026-07-16 | evidence |
| MiniMax M3 | MiniMax | 0.742 | 2026-06-01 | evidence |
| Muse Glimmer-30B | Meta | 0.755 | 2026-08-10 | evidence |
| Muse Spark 1.1 | Meta | 0.881 | 2026-07-09 | evidence |
| Nova 2 Lite | Amazon | 0.246 | 2025-12-02 | evidence |
| Qwen3.6 Plus | Qwen | 0.741 | 2026-04-02 | evidence |
| Qwen3.6-35B-A3B | Qwen | 0.628 | 2026-04-16 | evidence |
| Qwen3.7 Max | Qwen | 0.764 | 2026-05-19 | evidence |
| Qwen3.7-Plus | Qwen | 0.732 | 2026-05-31 | evidence |
| Seed 2.1 Pro | ByteDance | 0.838 | 2026-06-24 | evidence |
| Seed 2.1 Turbo | ByteDance | 0.803 | 2026-06-24 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.