Phare Debunking (Spanish)
Safety · 2025-03-27
Phare task testing ability to debunk false or misleading claims. This row is the Spanish language split.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-5.2 | 99.6 |
| GPT-5.5 | 99.5 |
| Sonnet 4.5 | 99.5 |
| Opus 4.5 | 99.5 |
| Haiku 4.5 | 99.3 |
| GPT-5.1 | 99.3 |
| Sonnet 4.6 | 99.1 |
| Opus 4.6 | 98.8 |
| GPT-5 | 98.5 |
| Claude 3.5 Sonnet (Oct 2024) | 98.4 |
| Sonnet 5 | 98.3 |
| Kimi K2.6 | 98.3 |
| Qwen3.7 Max Preview | 98.0 |
| Claude 3.7 Sonnet | 97.4 |
| GPT-4.1 | 97.2 |
| Kimi K2.5 | 97.2 |
| Qwen-Plus (2025-01-25) | 96.7 |
| DeepSeek-V4-Flash | 96.6 |
| Qwen3 Max (rolling alias) | 96.6 |
| Haiku 3.5 | 96.5 |
| Opus 4.1 | 96.5 |
| GPT-5 Mini | 96.5 |
| GPT-5 Nano | 96.3 |
| Grok 4.3 | 96.3 |
| Qwen3.7-Plus | 95.7 |
| DeepSeek-R1-0528 | 95.4 |
| Gemini 1.5 Pro | 95.4 |
| DeepSeek-V4-Pro | 95.3 |
| GPT-4o | 94.6 |
| gpt-oss-120b | 94.5 |
| Gemini 3.1 Pro Preview | 94.3 |
| Qwen3 8B | 93.8 |
| Gemma 4 31B IT | 93.4 |
| Gemini 2.0 Flash | 93.3 |
| Gemini 3.5 Flash | 92.9 |
| Llama 4 Maverick | 92.6 |
| Grok 3 Mini | 92.5 |
| Grok 4 | 92.1 |
| Gemini 3.1 Flash-Lite | 92.0 |
| GLM-5.2 | 91.5 |