OpenCompass Reasoning - Abductive - Chinese (CompassBench 2404): leaderboard
OpenCompass CompassBench 2404 edition (file dated 2024-05-09): abductive reasoning sub-task of the reasoning category, Chinese questions. CompassBench is OpenCompass's periodic evaluation of chat models on a private question set; each edition is a separate board. Edition 2405 republished 20 of these results for the same task set; those models appear on that board only. Score 0-100, higher is better.
Metric: Score (%). Source: rank.opencompass.org.cn. Saturation forecast: Estimated already saturated. 31 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4 Turbo (Preview) | 88.8 | #416 |
| 2 | OrionStar-Yi-34B-Chat | 77.5 | #872 |
| 3 | Yi Chat 34B | 69.4 | #950 |
| 4 | WizardLM-70B-V1.0 | 60.2 | #1219 |
| 5 | deepseek-llm-7B-chat | 51 | #1320 |
| 6 | Qwen-14B-Chat | 39.8 | #1161 |
| 7 | Baichuan2-7B-Chat | 34.7 | #1384 |
| 8 | vicuna-13B-v1.5-16k | 33.7 | #1417 |
| 9 | WizardLM-13B-V1.2 | 31.6 | #1399 |
| 10 | Qwen-72B-Chat | 29.6 | #835 |
| 11 | Llama 2 13B Chat | 28.6 | #1414 |
| 12 | Vicuna-7B-v1.5-16k | 27.6 | #1470 |
| 13 | Mistral 7B Instruct (v0.2) | 22.4 | #1311 |
| 14 | Qwen-7B Chat | 22.4 | #1353 |
| 15 | Llama 2 7B Chat | 16.3 | #1478 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-reasoning-abductive-chinese-compassbench-2404 · How It Works · Data refreshed daily, snapshot 2026-11-05.