Agentic Coding Benchmark

CWE-Bench (Resolve Rate)

Common Weakness Enumeration vulnerability-resolution agent benchmark. Model-card scores are the Resolve Rate over the collinear-ai/cwe-bench dataset (https://huggingface.co/datasets/collinear-ai/cwe-bench), reported via the cwe-bench-leaderboard HF space (https://huggingface.co/spaces/collinear-ai/cwe-bench-leaderboard).

Category

Agentic Coding

Score type

percent

Scale

100%

higher is better

Models tested

1

What is CWE-Bench (Resolve Rate)?

CWE-Bench (Resolve Rate) is a benchmark for how well a model works as an autonomous coding agent. The official description says: "Common Weakness Enumeration vulnerability-resolution agent benchmark". Below we describe in simple words what benchmarks of this type generally examine.

What does it test?

CWE-Bench (Resolve Rate) measures how well a model works as an autonomous coding agent. In detail, it checks:

  • fixing real, open bugs from real software repositories
  • planning multi-step changes, running commands and editing files by itself
  • passing the project's test suite after its changes
  • handling whole repositories, not just small isolated code snippets

How to read the score

The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better.

Official methodology note: Common Weakness Enumeration vulnerability-resolution agent benchmark. Model-card scores are the Resolve Rate over the collinear-ai/cwe-bench dataset (https://huggingface.co/datasets/collinear-ai/cwe-bench), reported via the cwe-bench-leaderboard HF space (https://huggingface.co/spaces/collinear-ai/cwe-bench-leaderboard).

Frequently asked questions

What is the CWE-Bench (Resolve Rate) benchmark?

CWE-Bench (Resolve Rate) is a benchmark for how well a model works as an autonomous coding agent. The official description says: 'Common Weakness Enumeration vulnerability-resolution agent benchmark'. Below we describe in simple words what benchmarks of this type generally examine.

What does CWE-Bench (Resolve Rate) test?

CWE-Bench (Resolve Rate) mainly measures how well a model works as an autonomous coding agent. Concretely, it checks: fixing real, open bugs from real software repositories; planning multi-step changes, running commands and editing files by itself; passing the project's test suite after its changes.

How is the CWE-Bench (Resolve Rate) score calculated?

The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better. The declared maximum score of this benchmark is 100.0.

Where do I get CWE-Bench (Resolve Rate) data and scores?

The official source is https://huggingface.co/datasets/collinear-ai/cwe-bench. The official benchmark page/reference is linked on this page. Every score on Open LLM Radar is also available as open data: download the scores JSON (http://open-llm-radar-api:8000/api/v1/benchmarks/cwe-bench-resolve-rate/scores) or pull it directly from the Open LLM Radar API with any HTTP client.

Model scores 1 model tested

Reset
# Model Score Date Source Verified
1 DeepSeek-V4.1-Flash-1017
DeepSeek
57%
17.10.2026 Report Verified

What now?

You've seen how models score on CWE-Bench (Resolve Rate) — here's what to do next.