Agentic Presentation Generation on 48 Business Briefs 1.0 (test)
0.794Overall QualityClaude Opus 4.6
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Claude Opus 4.62026.03 | 0.794 | 1 | 27.8 | 7.3 | 393.3 | |
| Llama 4 Scout2026.03 | 0.779 | 1 | 18.3 | 7.3 | 155.4 | |
| Claude Sonnet 4.62026.03 | 0.775 | 1 | 29.2 | 7.3 | 421.7 | |
| SlideRLBackbone=Qwen 7B, Fine-tuned=true, Training method=GRPO2026.03 | 0.724 | 0.9583 | 22.3 | 7 | 71.6 | |
| Base QwenBackbone=Qwen 7B2026.03 | 0.544 | 0.7083 | 19.1 | 5.2 | 43.8 | |
| GPT OSS 120BParameters=120B2026.03 | 0.249 | 0.3125 | 6.5 | 2.2 | 66.1 |