Galileo Agent - Banking TSQ: leaderboard
Task Success Quality score for Galileo banking workflows, measuring how completely and cleanly agents satisfy multi-step customer-service goals.
Metric: Task Success Quality (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Flash | 94 | #245 |
| 2 | GLM 4.5 Air | 94 | #392 |
| 3 | Grok 4 | 92 | #181 |
| 4 | Claude Sonnet 4 (20250514) | 90 | #213 |
| 5 | Kimi K2 | 89 | #243 |
| 6 | Gemini 2.5 Pro | 88 | #152 |
| 7 | Qwen 3 235B A22B 2507 Instruct | 88 | #300 |
| 8 | Qwen 3 235B A22B 2507 (Thinking) | 84 | #260 (Qwen 3 235B A22B 2507) |
| 9 | Gemini 2.5 Flash Lite | 82 | #422 |
| 10 | GPT-4.1 | 81 | #248 |
| 11 | GPT-4.1 Mini | 80 | #355 |
| 12 | DeepSeek V3 | 80 | #321 |
| 13 | Qwen 2.5 72B Instruct | 78 | #454 |
| 14 | Mistral Medium 3.1 | 74 | #408 |
| 15 | Mistral Small 3.2 | 73 | #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-tsq · How It Works · Data refreshed daily, snapshot 2026-10-20.