# ProgramBench leaderboard

Coding benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelscale.dev/leaderboards/benchmarks/coding-program-bench

ProgramBench: Can Language Models Rebuild Programs From Scratch?. Every model the catalog carries a published ProgramBench value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Coding |
| Measure | Cleanroom executable reimplementation |
| Tasks | 200 program reconstruction tasks |
| Difficulty | Full-repository software architecture |
| Published by | [ProgramBench: Can Language Models Rebuild Programs From Scratch?](https://programbench.com/static/paper.pdf) |

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.

## ProgramBench ranking

21 models with a published ProgramBench value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.

| Rank | Model | Provider | Cleanroom executable reimplementation | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [Claude Opus 5](https://modelscale.dev/models/claude-opus-5) | Anthropic | 93.0 | Observed 2026-11-21 · source benchlm:models |
| 2 | [Claude Opus 5.5](https://modelscale.dev/models/claude-opus-5-5) | Anthropic | 91.2 | Observed 2026-11-21 · source benchlm:models |
| 3 | [Claude Fable 5.1](https://modelscale.dev/models/claude-fable-5-1) | Anthropic | 87.6 | Observed 2026-11-21 · source benchlm:models |
| 4 | [Claude Fable 5.2](https://modelscale.dev/models/claude-fable-5-2) | Anthropic | 85.6 | Observed 2026-11-21 · source benchlm:models |
| 5 | [Step 5 Preview](https://modelscale.dev/models/step-5-preview) | StepFun | 80.5 | Observed 2026-11-21 · source benchlm:models |
| 6 | [Claude Sonnet 5.5](https://modelscale.dev/models/claude-sonnet-5-5) | Anthropic | 79.7 | Observed 2026-11-21 · source benchlm:models |
| 7 | [Kimi K3](https://modelscale.dev/models/kimi-3) | Moonshot AI | 77.8 | Observed 2026-11-21 · source benchlm:models |
| 8 | [GLM-5.2](https://modelscale.dev/models/glm-5-2) | Z.AI | 63.7 | Observed 2026-11-21 · source benchlm:models |
| 9 | [Muse Wonder Full](https://modelscale.dev/models/muse-wonder-full) | Meta | 59.0 | Observed 2026-11-21 · source benchlm:models |
| 10 | [Kimi K2.7 Code](https://modelscale.dev/models/kimi-k2-7-code) | Moonshot AI | 53.6 | Observed 2026-11-21 · source benchlm:models |
| 11 | [Muse Wonder](https://modelscale.dev/models/muse-wonder) | Meta | 50.0 | Observed 2026-11-21 · source benchlm:models |
| 12 | [Qwen4 Max](https://modelscale.dev/models/qwen4-max) | Alibaba | 31.5 | Observed 2026-11-21 · source benchlm:models |
| 13 | [Hy4](https://modelscale.dev/models/hy4) | Tencent | 30.0 | Observed 2026-11-21 · source benchlm:models |
| 14 | [ERNIE 5.5 Flash Preview](https://modelscale.dev/models/ernie-5-5-flash-preview) | Baidu | 29.5 | Observed 2026-11-21 · source benchlm:models |
| 15 | [DeepSeek V4.1 Flash 1017](https://modelscale.dev/models/deepseek-v4-1-flash) | DeepSeek | 29.1 | Observed 2026-11-21 · source benchlm:models |
| 16 | [MiMo-V2.6-Pro](https://modelscale.dev/models/mimo-v2-6-pro) | Xiaomi | 26.5 | Observed 2026-11-21 · source benchlm:models |
| 17 | [MiMo-V2.6-Flash](https://modelscale.dev/models/mimo-v2-6-flash) | Xiaomi | 26.0 | Observed 2026-11-21 · source benchlm:models |
| 18 | [LongCat-2.5](https://modelscale.dev/models/longcat-2-5) | Meituan | 25.5 | Observed 2026-11-21 · source benchlm:models |
| 19 | [GLM-5.3](https://modelscale.dev/models/glm-5-3) | Z.AI | 19.0 | Observed 2026-11-21 · source benchlm:models |
| 20 | [Hy4 preview](https://modelscale.dev/models/hy4-preview) | Tencent | 17.5 | Observed 2026-11-21 · source benchlm:models |
| 21 | [Hy4 flash](https://modelscale.dev/models/hy4-flash) | Tencent | 16.0 | Observed 2026-11-21 · source benchlm:models |

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
