New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
MCP-Mark benchmark
AI model leaderboard for the MCP-Mark benchmark. Compare how large language models score on MCP-Mark, see the full ranking, and understand what this AI benchmark measures. GPT-5 currently leads with 50.9. Benchmark for MCP-tool calling correctness on isolated, well-specified single-server tasks.
Leaderboard
| # | Model | Organization | Score | Variant | Source |
|---|---|---|---|---|---|
| #1 | GPT-5 | OpenAI | 50.9 | cited-deepseek-v3-2 | official ↗ |
| #2 | Gemini 3 Pro | Google DeepMind | 43.1 | cited-deepseek-v3-2 | official ↗ |
| #3 | DeepSeek-V3.2 | DeepSeek | 38 | thinking | official ↗ |
| #4 | Claude Sonnet 4.5 | Anthropic | 33.3 | cited-deepseek-v3-2 | official ↗ |
| #5 | MiniMax-M2 | MiniMax | 24.4 | cited-deepseek-v3-2 | official ↗ |
| #6 | Kimi K2 Thinking | Moonshot | 20.4 | thinking-cited-deepseek-v3-2 | official ↗ |
Frequently asked questions about MCP-Mark
What is the MCP-Mark benchmark?
Benchmark for MCP-tool calling correctness on isolated, well-specified single-server tasks.
How is the MCP-Mark benchmark scored?
MCP-Mark is scored using the accuracy metric, where a higher score is better. GenAIList aggregates reported scores from model providers and papers into a single ranked leaderboard.
Which AI model scores highest on MCP-Mark?
As of the latest reported scores on GenAIList, GPT-5 achieves the highest result on MCP-Mark with a score of 50.9.
Is a higher MCP-Mark score better?
Yes. On MCP-Mark a higher score indicates better performance, so models near the top of the leaderboard are the strongest.