reasoning benchmark
FrontierCode
FrontierCode is Cognition's coding evaluation that tests whether models can pass difficult coding tasks while meeting the standards of high-quality production codebases. The Diamond subset contains the hardest problems.
Updated Aug 11, 2026
Models3
Model coverage3
MetricScore
EvidenceB
FrontierCode Ranking
Higher score ranks better on this benchmark.
| 01 | AN | 53.4% | 100.0% | 3 | C | |
| 02 | AN | 46.3% | 50.0% | 3 | C | |
| 03 | AN | 38.8% | 0.0% | 3 | C |
FrontierCode Score Distribution
A closer view of the leading scores on this benchmark.
FrontierCode Highlights
The leading models and scores on this benchmark.
What is FrontierCode?
What FrontierCode measures and how its scores work.
FrontierCode is Cognition's coding evaluation that tests whether models can pass difficult coding tasks while meeting the standards of high-quality production codebases. The Diamond subset contains the hardest problems.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- FrontierCode
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- frontiercode|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about FrontierCode.
Which model scores highest on FrontierCode?
Claude Opus 5 is currently ranked first with 53.4%.
What does FrontierCode measure?
FrontierCode is Cognition's coding evaluation that tests whether models can pass difficult coding tasks while meeting the standards of high-quality production codebases. The Diamond subset contains the hardest problems.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
3 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.