Benchmark·Multimodal reasoning
MMMU-Benchmark
MMMU-Pro
A multimodal reasoning benchmark used in frontier model evaluation.
23 curated resources
AI Agent Benchmark Resources
A curated library of public AI agent benchmarks, leaderboards, papers, tools, articles, and reports with editorial judgment.
Showing 13-23 of 23 resourcesPage 2 of 2
Open resource
Benchmark·Math reasoning
OpenAI
MATH-500
A 500-problem math benchmark built from the held-out MATH test split.
Benchmark·Scientific coding
SciCode
SciCode
A benchmark for coding solutions to realistic scientific research problems.
Benchmark·Terminal agents
Harbor Framework
Terminal-Bench v2.1
An evaluation for coding agents that use terminal tools to complete tasks.
Leaderboard·General AI assistants
GAIA Benchmark
GAIA Leaderboard
A leaderboard for a benchmark focused on next-generation assistants with augmented capabilities such as tooling and search.
Leaderboard·Function calling and tools
Gorilla / Berkeley
Berkeley Function Calling Leaderboard
A leaderboard for evaluating how well models select, format, and execute tool or function calls.
Paper·Multi-environment agents
THUDM
AgentBench
A benchmark proposal for evaluating LLMs as agents across multiple environments and task categories.
Tool·Agent testing and monitoring
LangChain
LangSmith Evaluation
A platform for datasets, evaluators, experiments, tracing, and production quality monitoring for LLM and agent applications.
Tool·Benchmark discovery
Koutian Wu
Benchmark Radar
A public dashboard and dataset for finding benchmark signals and leaderboard snapshots.
Article·Evaluation and judgment
AIAgentBenchmark
Why AI Agent Benchmark Matters
A field note on why benchmarks matter when agents move from demos to real work and the cost of judgment becomes part of the workflow.
Article·Agent system design
Anthropic Engineering
Building Effective Agents
An engineering article on patterns for building agent systems, from workflows to more autonomous agents.
Resource does not match your workflow?