Phare Misinformation (Spanish)
Safety · 2025-02-19
Phare task testing resistance to misinformation generation or endorsement. This row is the Spanish language split.
Top models (higher is better)
| Model | Score |
|---|---|
| Haiku 4.5 | 91.2 |
| Claude 3.7 Sonnet | 87.8 |
| Sonnet 5 | 84.3 |
| DeepSeek-V4-Flash | 83.7 |
| GPT-5 Nano | 83.0 |
| Opus 4.5 | 81.0 |
| Claude 3.5 Sonnet (Oct 2024) | 79.6 |
| Sonnet 4.5 | 79.6 |
| Magistral Medium 1.2 | 79.6 |
| GPT-4.1 Nano | 78.2 |
| Qwen-Plus (2025-01-25) | 75.5 |
| Gemini 1.5 Pro | 74.8 |
| GPT-5 Mini | 74.8 |
| Opus 4.1 | 74.2 |
| Opus 4.6 | 74.2 |
| Haiku 3.5 | 70.8 |
| Llama 3.1 8B Instruct | 69.4 |
| Mistral Large 2 (Instruct 2407) | 69.4 |
| Gemini 2.5 Flash | 68.0 |
| Gemini 2.5 Flash-Lite | 66.7 |
| Gemini 3.1 Pro Preview | 66.7 |
| Llama 3.1 405B Instruct | 66.7 |
| Sonnet 4.6 | 66.0 |
| Kimi K2.6 | 66.0 |
| GPT-4o | 65.3 |
| Grok 4.3 | 62.6 |
| Qwen3.7 Max Preview | 61.9 |
| Gemini 3.1 Flash-Lite | 60.5 |
| Command A | 59.9 |
| Llama 4 Maverick | 59.9 |
| Gemma 4 31B IT | 59.2 |
| Gemini 3 Pro Preview | 57.8 |
| GPT-5.1 | 57.1 |
| Grok 3 Mini | 57.1 |
| Gemini 2.5 Pro | 56.5 |
| Mistral Small 3.1 24B Instruct 2503 | 56.5 |
| Qwen2.5-Max | 56.5 |
| Grok 4 | 53.7 |
| Gemini 2.0 Flash | 53.1 |
| Kimi K2.5 | 53.1 |