FACTS Benchmark Suite
FACTS Benchmark Suite scores LLM factuality across Grounding, Multimodal, Parametric and Search; Gemini 3 Pro leads at 68.8% overall.
Extends the 2024 FACTS Grounding benchmark into four parts, with Search and Parametric error rates dropping 55% and 35% from Gemini 2.5 Pro to 3 Pro. Multimodal scores were the lowest.
- Date
- Tuesday, 9 December 2025
- Lab
- Google DeepMind
- Kind
- paper
- Access
- research preview
Figures
| Measure | Value | Measured by |
|---|---|---|
| FACTS Score (Gemini 3 Pro) | 68.8% overall of four sub-benchmarks; Google-run | company |
Built and run by Google with Kaggle.
Sources
This record was checked and corrected against its sources on 6 October 2026. How we check