SAHARA - Text Generation Cluster: leaderboard
Metric: Cluster score (mean of task scores). Source: huggingface.co. 47 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 18.24 |
| 2 | Gemini 2.5 Pro | 15.66 |
| 3 | GPT-5.6 Luna | 13.54 |
| 4 | Qwen 3.6 35B A3B | 12.42 |
| 5 | Claude Sonnet 4 | 12.04 |
| 6 | Qwen 3.6 27B | 11.96 |
| 7 | Qwen 3.5 27B | 11.73 |
| 8 | Llama 4 Scout Instruct | 11.45 |
| 9 | Gemini 2.5 Flash | 11.18 |
| 10 | Llama 3.1 70B Instruct | 11.15 |
| 11 | GPT-5 | 11.1 |
| 12 | Qwen 3.5 4B | 11.08 |
| 13 | GPT-4.1 | 11.06 |
| 14 | DeepSeek R1 Distill Llama 70B | 10.91 |
| 15 | Command-R+ (08-2024) | 10.85 |
Interactive version: theaggregate.ai/benchmark?slug=sahara-text-generation-cluster · How It Works · Data refreshed daily, snapshot 2026-09-19.