Subjective Guidance Quality on HowTo
3.93Subjective Quality ScoreLlama 4-Maverick (PWR)
Evaluation Results
| Method | Links | |
|---|---|---|
| Llama 4-Maverick (PWR)Config=PWR2026.06 | 3.93 | |
| Llama 4-MaverickConfig=FT2026.06 | 2.93 | |
| Qwen3.6-VL-27B (PWR)Config=PWR2026.06 | 2.67 | |
| Qwen3.6-VL-27BConfig=ZS2026.06 | 2.62 | |
| Gemini 3.1 ProConfig=ZS2026.06 | 2.41 | |
| Qwen3-VL-235BConfig=ZS2026.06 | 2.32 | |
| Claude Opus 4.6Config=ZS2026.06 | 2.09 | |
| Qwen3.6-VL-27BConfig=FT2026.06 | 2.08 | |
| Llama 4-MaverickConfig=ZS2026.06 | 1.97 |