Galileo Agent - Banking Accuracy: leaderboard
Galileo agent benchmark slice for banking workflows, measuring whether agents complete customer-service goals correctly across account, card, payment, and branch-assistance tasks.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around January 2027. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 | 60 | #248 |
| 2 | Kimi K2 | 58 | #243 |
| 3 | Claude Sonnet 4 (20250514) | 58 | #213 |
| 4 | Mistral Medium 3.1 | 57 | #408 |
| 5 | GPT-4.1 Mini | 56 | #355 |
| 6 | GLM 4.5 Air | 49 | #392 |
| 7 | Gemini 2.5 Flash | 48 | #245 |
| 8 | Qwen 2.5 72B Instruct | 48 | #454 |
| 9 | Gemini 2.5 Pro | 45 | #152 |
| 10 | Gemini 2.5 Flash Lite | 45 | #422 |
| 11 | Qwen 3 235B A22B 2507 Instruct | 44 | #300 |
| 12 | Qwen 3 235B A22B 2507 (Thinking) | 42 | #260 (Qwen 3 235B A22B 2507) |
| 13 | GPT-4.1 Nano | 40 | #731 |
| 14 | DeepSeek V3 | 38 | #321 |
| 15 | Mistral Small 3.2 | 37 | #605 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=galileo-agent-banking-accuracy · How It Works · Data refreshed daily, snapshot 2026-10-20.