TextClass Benchmark: leaderboard
TextClass Benchmark evaluates LLMs and transformers for social-science text classification across multiple domains and languages, reporting domain-specific Elo leaderboards and a weighted Meta-Elo aggregate.
Metric: Meta-Elo (imported). Source: github.com. Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 108 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-05-13) | 1825.22 | #325 |
| 2 | Gemini 1.5 Pro | 1782.7 | #445 |
| 3 | GPT-4 Turbo | 1781.47 | #404 |
| 4 | O1 (2024-12-17) | 1768.81 | #149 |
| 5 | GPT-4.5 | 1767.86 | #274 |
| 6 | Grok 2 (1212) | 1758.36 | #496 |
| 7 | Llama 3.1 405B | 1755.81 | #466 |
| 8 | GPT-4 (0613) | 1747.59 | #491 |
| 9 | Llama 3.3 70B | 1746.41 | #579 |
| 10 | Grok Beta | 1741.94 | #487 |
| 11 | DeepSeek V3 | 1732.54 | #321 |
| 12 | Mistral Large 2 (Nov) Instruct (2411) | 1720.36 | #465 |
| 13 | DeepSeek R1 | 1718.73 | #251 |
| 14 | Gemini 2.0 Flash | 1701.95 | #342 |
| 15 | pixtral-large-2411 | 1697.33 | #478 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=textclass-benchmark · How It Works · Data refreshed daily, snapshot 2026-10-20.