agents benchmark
MCP-Mark
MCP-Mark evaluates LLMs on their ability to use Model Context Protocol (MCP) tools effectively, testing tool discovery, selection, invocation, and result interpretation across diverse MCP server scenarios.
Updated Aug 7, 2026
Published models8
Registry coverage8
MetricScore
EvidenceB
MCP-Mark leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | MA | 81.1% | 100.0% | 8 | C | |
| 2 | AC | 60.8% | 85.7% | 8 | C | |
| 3 | AC | 58.7% | 71.4% | 8 | C | |
| 4 | MA | 55.9% | 57.1% | 8 | C | |
| 5 | AC | 48.2% | 42.9% | 8 | C | |
| 6 | AC | 46.1% | 28.6% | 8 | C | |
| 7 | DE | 38.0% | 14.3% | 8 | C | |
| 8 | AC | 37.0% | 0.0% | 8 | C |
Score distribution
Top published rows on the benchmark's original scale.
MCP-Mark highlights
The top published results on this benchmark's own scale.
What is MCP-Mark?
Definition and scoring fields from the benchmark registry.
MCP-Mark evaluates LLMs on their ability to use Model Context Protocol (MCP) tools effectively, testing tool discovery, selection, invocation, and result interpretation across diverse MCP server scenarios.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- MCP-Mark
- Modality
- text
- Primary category
- agents
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- mcp-mark|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about MCP-Mark.
Which model scores highest on MCP-Mark?
Kimi K2.7 Code is currently ranked first with 81.1%.
What does MCP-Mark measure?
MCP-Mark evaluates LLMs on their ability to use Model Context Protocol (MCP) tools effectively, testing tool discovery, selection, invocation, and result interpretation across diverse MCP server scenarios.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
8 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.