AI benchmarks and leaderboards are different things
Learn the difference between a benchmark, an evaluation run and a leaderboard, and use each to answer the right question.
LLM Stats
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.
Evaluation context and evidence
- Source
- LLM Stats
- Publisher
- Not recorded
- Benchmark release
- Not recorded
- Collected
- 2026-08-17
- Modality
- text
These are source-reported observations, not independent reruns by this site. Check the task version, evaluation protocol, unit and reporting date before comparing models. Results from different setups are not a universal ranking.
Original sources
Reported results 6
Shown in source observation order, not ranked. Dates retain the precision provided by the source.
| Model | Value | Unit | Protocol | Reported date and basis | Source |
|---|---|---|---|---|---|
| Kimi K3 | 0.778 | Not recorded | Not recorded | 2026-07-16 · model_announcement | LLM Stats |
| GLM-5.2 | 0.637 | Not recorded | Not recorded | 2026-06-16 · model_announcement | LLM Stats |
| Kimi K2.7 Code | 0.536 | Not recorded | Not recorded | 2026-06-12 · model_announcement | LLM Stats |
| Seed 2.1 Pro | 0.503 | Not recorded | Not recorded | 2026-06-24 · model_announcement | LLM Stats |
| Seed 2.1 Turbo | 0.494 | Not recorded | Not recorded | 2026-06-24 · model_announcement | LLM Stats |
| GLM-5.3 | 0.19 | Not recorded | Not recorded | 2026-08-14 · model_announcement | LLM Stats |