Agentic Output Verification on GAIA Plan Human-annotated (test)
93Accuracyft-4.1-nano
Evaluation results
| Method | Links | |
|---|---|---|
| ft-4.1-nanoCategory=Fine-tuned Model, Training Method=BARRED2026.04 | 93 | |
| GPT-5-miniCategory=LLM-as-a-Judge baseline2026.04 | 92 | |
| ft-Qwen14BCategory=Fine-tuned Model, Training Method=BARRED2026.04 | 92 | |
| OSS-Safeguard-20BCategory=Generic Guardrail Model2026.04 | 87 | |
| ft-Qwen3BCategory=Fine-tuned Model, Training Method=BARRED2026.04 | 87 | |
| GPT-4.1-miniCategory=LLM-as-a-Judge baseline2026.04 | 83 | |
| GPT-4.1Category=LLM-as-a-Judge baseline2026.04 | 80 | |
| GliderCategory=Generic Guardrail Model2026.04 | 65 | |
| GPT-4.1-nanoCategory=LLM-as-a-Judge baseline2026.04 | 52 | |
| Qwen14BCategory=LLM-as-a-Judge baseline2026.04 | 50 |