ProgramBench: Can Language Models Rebuild Programs From Scratch? (ProgramBench)
We mirror this table; we do not rank on it.
A cleanroom software-engineering benchmark where agents receive only a compiled executable and documentation, then must architect and implement a complete codebase that reproduces the original program's behavior.
Almost resolved rate on ProgramBench — Sep. 9, 2026
We mirror the published almost resolved rate view for ProgramBench. Claude Opus 5 leads the public snapshot at 37.0%, followed by Claude Opus 4.8 (16.5%) and GPT-5.6 Sol (15.5%). We do not use these results to rank models overall.
Claude Opus 5
Anthropic
Claude Opus 4.8
Anthropic
GPT-5.6 Sol
OpenAI
21 modelsCodingCurrentDisplay onlyUpdated Sep. 9, 2026
Almost resolved rate table (21 models)
ScoreHow ProgramBench is shown here
BenchLM mirrors the public ProgramBench leaderboard from Sep. 9, 2026. The official benchmark uses mini-SWE-agent on 200 cleanroom program-reconstruction tasks and reports both fully resolved and almost-resolved rates.
BenchLM displays the published almost-resolved rate as the visible score because the primary fully resolved metric is still very low and less differentiating. Almost resolved means the agent passed at least 95% of the hidden behavioral tests for a task. ProgramBench remains display only on BenchLM and is excluded from weighted coding and overall rankings until public model coverage broadens.
Snapshot
The published ProgramBench snapshot places Claude Opus 5 first at 37.0%. The third row is 21.5 points behind. The broader top-10 range is 34.0 points, so the table still separates the published systems.
21 models have been evaluated on ProgramBench. The benchmark falls in the Coding category. This category carries a 20% weight in BenchLM.ai's overall scoring system. ProgramBench is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About ProgramBench
Year
2026
Tasks
200 program reconstruction tasks
Format
Cleanroom executable reimplementation
Difficulty
Full-repository software architecture
ProgramBench turns open-source projects into cleanroom reconstruction tasks. Each task starts from an execute-only binary and usage documentation, with no source code, internet, decompilation, or prescribed skeleton. Evaluation uses hidden behavioral tests generated through agent-driven fuzzing. BenchLM shows ProgramBench as display-only because all current public rows are tied at 0% fully resolved and the visible score is the auxiliary almost-resolved metric.
Freshness and provenance
Version
ProgramBench 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public task metadata and hidden behavioral tests
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does ProgramBench measure?
A cleanroom software-engineering benchmark where agents receive only a compiled executable and documentation, then must architect and implement a complete codebase that reproduces the original program's behavior.
Which model leads the published ProgramBench snapshot?
Claude Opus 5 currently leads the published ProgramBench snapshot with 37.0% almost resolved rate. BenchLM shows this benchmark for display only and does not use it in overall rankings.
How many models are evaluated on ProgramBench?
The Sep. 9, 2026 snapshot contains 21 AI models.
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 2,000+ readers.
One email each week. Unsubscribe anytime.