ActiveCoding
FrontierCode
Private maintainer-authored repository tasks graded for mergeability: correctness, tests, scope discipline, style, and code quality.
Top Benchmarks & Trusted AI Evaluation Directory
Comprehensive directory of 63+ LLM benchmarks with honest status labels, contamination risk assessments, and vendor vs independent score provenance as of September 2026.
Capability Domains
Status
Contamination
Sort By
Displaying 16 of 16 benchmarks (filtered from 63)
ActiveCoding
SWE-bench Pro
Long-horizon repository tasks across public, held-out, and commercial codebases, with a continuing public model leaderboard.
ActiveCoding
Terminal-Bench
Eighty-nine realistic terminal tasks graded in containers; version 2.1 repairs 2.0's dependency, resource, and specification defects.
Nearing SaturationCoding
MBPP+
A cleaned 378-task MBPP evaluation with roughly 35 times more tests for stricter Python functional correctness.
SaturatedCoding
LiveCodeBench
Continuously updated contest problems for time-windowed code generation, execution, test prediction, and self-repair evaluation.
SaturatedCoding
MBPP
Mostly Basic Programming Problems: 974 crowd-sourced Python synthesis tasks intended for entry-level programmers.
SaturatedCoding
SWE-bench Verified
A 500-task, engineer-reviewed subset of SWE-bench that became the standard repository-level coding-agent test.
SaturatedCoding
HumanEval
OpenAI's 164 hand-written Python function-synthesis tasks, graded by executing generated completions against unit tests.
SaturatedCoding
HumanEval+
HumanEval's 164 Python tasks re-evaluated with roughly 80 times more tests to expose plausible but incorrect programs.
DeprecatedCoding
SWE-Lancer
Current 198-task offline patch benchmark derived from paid Upwork work, with historical mixed implementation and manager-task variants.
DeprecatedCoding
CodeXGLUE
A broad code-intelligence suite spanning understanding, retrieval, completion, translation, repair, generation, and summarization.
DeprecatedCoding
Mercury
Python code-synthesis benchmark that rewards both functional correctness and runtime efficiency against real solution distributions.
DeprecatedCoding
MultiPL-E
Parallel HumanEval and MBPP translations for comparing code generation across many programming languages.
DeprecatedCoding
SWE-bench
Real GitHub issues from 12 Python repositories, solved by generating patches that must pass repository tests.
DeprecatedCoding
CodeContests
DeepMind's competitive-programming corpus with temporally split problems, human submissions, and generated correctness tests.
DeprecatedCoding
CyberSecEval
Meta's evolving suite for insecure code, cyberattack assistance, prompt injection, exploitation, SOC analysis, and autopatching.