Agentic Reasoning on TauBench V2
76Airline ScoreGemma 4 26B-A4B
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Gemma 4 26B-A4BModel size=26B-A4B, Thinking mode=true2026.07 | 76 | 85.5 | 43 | — | |
| Gemma 4 31BModel size=31B, Thinking mode=true2026.07 | 75 | 86.4 | 69.3 | — | |
| Gemma 4 12BModel size=12B, Thinking mode=true2026.07 | 75 | 77.6 | 54.4 | — | |
| Qwen3.5-122B-A10B2026.04 | 66 | 62.6 | 95 | 74.53 | |
| Nemotron 3 Super2026.04 | 56.25 | 62.83 | 64.36 | 61.15 | |
| Gemma 4 E4BModel size=E4B, Thinking mode=true2026.07 | 52 | 67.1 | 18.4 | — | |
| GPT-OSS-120B2026.04 | 49.2 | 67.8 | 66 | 61 | |
| Gemma 3 27BModel size=27B, Thinking mode=false2026.07 | 39 | 6.6 | 3.1 | — | |
| Gemma 4 E2BModel size=E2B, Thinking mode=true2026.07 | 31 | 34.6 | 19.7 | — |