agents benchmark
Program Bench
Program Bench evaluates code-generation agents by asking them to recreate a program's behavior from only a compiled binary and documentation. It spans 200 tasks from small CLI tools to large systems such as FFmpeg and SQLite, with submissions judged against more than 248,000 fuzz-generated behavioral tests.
Updated Aug 11, 2026
Models5
Model coverage5
MetricScore
EvidenceB
Program Bench Ranking
Higher score ranks better on this benchmark.
| 1 | MA | 77.8% | 100.0% | 5 | C | |
| 2 | ZA | 63.7% | 75.0% | 5 | C | |
| 3 | MA | 53.6% | 50.0% | 5 | C | |
| 4 | BY | 50.3% | 25.0% | 5 | C | |
| 5 | BY | 49.4% | 0.0% | 5 | C |
Program Bench Score Distribution
A closer view of the leading scores on this benchmark.
Program Bench Highlights
The leading models and scores on this benchmark.
What is Program Bench?
What Program Bench measures and how its scores work.
Program Bench evaluates code-generation agents by asking them to recreate a program's behavior from only a compiled binary and documentation. It spans 200 tasks from small CLI tools to large systems such as FFmpeg and SQLite, with submissions judged against more than 248,000 fuzz-generated behavioral tests.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- Program Bench
- Modality
- text
- Primary category
- agents
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- program-bench|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about Program Bench.
Which model scores highest on Program Bench?
Kimi K3 is currently ranked first with 77.8%.
What does Program Bench measure?
Program Bench evaluates code-generation agents by asking them to recreate a program's behavior from only a compiled binary and documentation. It spans 200 tasks from small CLI tools to large systems such as FFmpeg and SQLite, with submissions judged against more than 248,000 fuzz-generated behavioral tests.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
5 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.