ReLE - Coding: leaderboard
ReLE Chinese LLM benchmark (NoneLinear, formerly CLiB), a continuously updated Chinese-language evaluation run through each model's API. Coding domain, the mean of ReLE's own LiveCodeBench and Terminal-Bench 2.0 runs; read from the repository's generated coding board because the site API omits this domain. Accuracy 0-100, higher is better.
Metric: Accuracy (%). Source: raw.githubusercontent.com. Saturation forecast: Around March 2027. 68 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 5.5 | 78 | #14 |
| 2 | Claude Opus 5.5 | 76.3 | #3 |
| 3 | Qwen 3.6 Max Preview | 75.7 | #91 |
| 4 | GPT-5.6 Pro Sol | 75 | #21 |
| 5 | GPT-6 Astra | 74.8 | #6 |
| 6 | Gemini 3.7 Flash | 74.8 | #24 |
| 7 | Gemini 3.1 Pro (Preview) | 74.2 | #64 |
| 8 | Gemini 3.5 Flash | 74.2 | #66 |
| 9 | Grok 4.6 | 74.2 | #29 |
| 10 | Qwen 3.7 Max | 73.3 | #85 |
| 11 | Grok 4.7 | 73.2 | #26 |
| 12 | Gemini 3.8 Flash | 72.7 | #30 |
| 13 | GPT-6 Sol | 72.7 | #19 |
| 14 | Claude Opus 5 | 72.5 | #16 |
| 15 | DeepSeek V4.1 Flash | 72.5 | #37 |
No result here: #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=rele-coding · How It Works · Data refreshed daily, snapshot 2026-11-05.