# PaperBench 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://modelchoice.dev/leaderboards/benchmarks/coding-paper-bench

Every model the catalog carries a published PaperBench value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Coding |
| Measure | Long-horizon agent evaluation |
| Tasks | AI research-paper reproduction |
| Difficulty | Frontier autonomous research and engineering |
| Published by | [Qwen3.8-Max: A New Bar for Coding and Cowork](https://qwen.ai/blog?id=qwen3.8) |

A research-reproduction benchmark that asks agents to recreate the contributions of AI papers from the paper alone.

## PaperBench ranking

2 models with a published PaperBench 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 | Long-horizon agent evaluation | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [Qwen3.8 Max](https://modelchoice.dev/models/qwen3-8-max) | Alibaba | 93.0 | Observed 2026-11-21 · source benchlm:models |
| 2 | [LongCat-2.5](https://modelchoice.dev/models/longcat-2-5) | Meituan | 82.1 | 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.
