Coding Benchmarks — Top AI Models

Benchmarks in this category

• SWE-Bench Verified — Resolve real GitHub issues end-to-end with passing test suite.
• HumanEval+ — Function-level code synthesis with extended hidden tests.
• LiveCodeBench — Fresh competitive programming problems with timestamped contamination guard.
• Codeforces ELO — Competitive programming ELO equivalent for model submissions.
• SWE-Bench Multimodal — Fix UI bugs that require screenshot understanding alongside code.
• SWE-Bench Lite — 300-task subset of SWE-Bench for faster eval.
• MBPP+ — Mostly Basic Python Problems with adversarial extended tests.
• BigCodeBench — Function-level synthesis requiring practical library use.
• CRUXEval — Code reasoning over input prediction and output prediction.
• HumanEval-X — Multilingual HumanEval across Python, JS, Java, C++, Go.
• MultiPL-E — HumanEval & MBPP translated into 18 programming languages.
• Spider 2.0 — Real-world text-to-SQL across enterprise warehouses.
• DS-1000 — Data-science code synthesis across NumPy, Pandas, PyTorch, scikit-learn.
• RepoBench — Repository-level code completion across long context.
• Aider Polyglot — Aider's polyglot editing benchmark across 6 languages and 225 Exercism tasks.
• Terminal-Bench — Agentic command-line task suite (Stanford / Anthropic).

Continue your research

Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Coding Benchmarks — Top AI Models.

• Browse models
• Review providers
• Inspect benchmarks
• Explore workflows
• Learn AI concepts
• Read decision guides
• Review trust and methodology
• Start a decision path