Behavior Tree Synthesis on 10 Real-world User Requests 1.0 (test)
100TSR (%)Gemini 3 Flash
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Gemini 3 FlashProtocol=one-shot2026.04 | 100 | 100 | 0 | 6.2 | 33.6 | 0.56 | 52.2 | 100 | — | |
| Pixtral-12BAblation ID=A12026.04 | 100 | 100 | 0 | 6 | 23.1 | 0.59 | 30.1 | 100 | 61.85 | |
| Pixtral-12BAblation ID=A92026.04 | 90 | 100 | 0 | 6 | 22.1 | 0.54 | 35.8 | 100 | 60.12 | |
| GPT-5Protocol=one-shot2026.04 | 80 | 80 | 0 | 6.2 | 40.4 | 0.58 | 64 | 80 | — | |
| Pixtral-12BAblation ID=A102026.04 | 80 | 100 | 10 | 6 | 25.1 | 0.58 | 38.4 | 90 | 52.24 | |
| Pixtral-12BAblation ID=A52026.04 | 70 | 100 | 30 | 6 | 21.8 | 0.48 | 32.1 | 90 | 54.42 | |
| Pixtral-12BAblation ID=A172026.04 | 70 | 100 | 0 | 6 | 21.4 | 0.51 | 34.1 | 100 | 59.88 | |
| Pixtral-12BAblation ID=A112026.04 | 60 | 90 | 20 | 5.4 | 19.2 | 0.42 | 28.1 | 100 | 54.39 | |
| Pixtral-12BAblation ID=A162026.04 | 50 | 90 | 10 | 4.8 | 15.6 | 0.38 | 24.5 | 80 | 58.21 | |
| Pixtral-12BAblation ID=A22026.04 | 40 | 100 | 60 | 6 | 12.8 | 0.3 | 33.4 | 100 | 58.45 | |
| Pixtral-12BAblation ID=A82026.04 | 40 | 100 | 60 | 6 | 12.8 | 0.3 | 33.4 | 100 | 61.64 | |
| Pixtral-12BAblation ID=A72026.04 | 30 | 40 | 0 | 2.1 | 6.7 | 0.2 | 10.1 | 30 | 40.54 | |
| Pixtral-12BAblation ID=A122026.04 | 30 | 70 | 30 | 4.2 | 14.5 | 0.31 | 19.8 | 80 | 54.45 | |
| Pixtral-12BAblation ID=A42026.04 | 20 | 50 | 0 | 3.4 | 11.2 | 0.28 | 15.6 | 100 | 54.66 | |
| Pixtral-12BAblation ID=A142026.04 | 20 | 40 | 10 | 2.4 | 7.1 | 0.18 | 12.4 | 40 | 58.27 | |
| Pixtral-12BAblation ID=A32026.04 | 10 | 10 | 0 | 0.6 | 1.9 | 0.06 | 3 | 10 | 53.86 | |
| Pixtral-12BAblation ID=A152026.04 | 10 | 40 | 20 | 1.8 | 5.2 | 0.12 | 8.4 | 20 | 55.45 | |
| GPT-5Protocol=zero-shot2026.04 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | — | |
| Gemini 3 FlashProtocol=zero-shot2026.04 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | — | |
| Pixtral-12BProtocol=zero-shot2026.04 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 52.8 | |
| Pixtral-12BAblation ID=A62026.04 | 0 | 10 | 0 | 0.8 | 2.1 | 0.04 | 3.5 | 0 | 18.42 | |
| Pixtral-12BAblation ID=A132026.04 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 309.71 |