0

MCP Universe

MCP Universe environment for evaluating LLMs in wide range of tasks with MCP server

Domain
rl-env
License
apache-2.0
Published
Oct 2025

Cite

Notes

Only stored in your browser.

Attribution

Leaderboard scores
prime-hub
Attribution policy →

Top score 92.7% by GPT-5 - 2 models reporting (2 frontier)

Score history

2
25%44%63%81%100%Apr 25May 25Jun 25Jul 25Aug 25GPT-4.1GPT-5

Top models

2
MCP UniverseBar chart with 2 bars. Highest value: GPT-5 at 92.7.
2 models

Related tools

1
View all

Implementations, trainers, datasets and scaffolds linked to this eval.

FAQ

What is MCP Universe?
MCP Universe environment for evaluating LLMs in wide range of tasks with MCP server
What is the current top score on MCP Universe?
The top reported score is 92.7% by GPT-5, across 2 models reporting (2 from frontier labs).
How can a model improve its MCP Universe score?
Tools linked to MCP Universe on Sophon include MCP Universe RL Env (Prime Intellect) - RL environments, datasets, and scaffolds that target this eval.
What license is MCP Universe under?
MCP Universe is available under apache-2.0.