FinBen - FinText: leaderboard
Financial textual analysis tasks evaluating sentiment classification and text categorization across financial news, analyst reports, and social media posts.
Metric: Normalized Score. Source: huggingface.co. Saturation forecast: Estimated already saturated. 25 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4 | 60 | #438 |
| 2 | Qwen 2.5 32B Instruct | 55.21 | #498 |
| 3 | Llama 3 70B Instruct | 44.57 | #634 |
| 4 | GPT-4.5 (Preview) | 40.67 | #244 |
| 5 | Qwen 2.5 72B Instruct | 38.77 | #454 |
| 6 | GPT-4o | 31.3 | #343 |
| 7 | GPT-3.5 Turbo (0125) | 30.32 | #984 |
| 8 | GPT-4o Mini | 30.17 | #599 |
| 9 | Llama 3.1 8B Instruct | 20.74 | #1035 |
| 10 | Gemma 2 27B (IT) | 18.49 | #645 |
| 11 | Gemma 2 2B (IT) | 16.03 | #1341 |
| 12 | Llama 3 8B Instruct | 13.25 | #1129 |
| 13 | Qwen 2.5 7B Instruct | 13.2 | #862 |
| 14 | Gemma 2 9B (IT) | 4.55 | #788 |
| 15 | DeepSeek V3 Chat | 0.25 | #447 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=finben-fintext · How It Works · Data refreshed daily, snapshot 2026-10-20.