Phare Tools Reliability (Spanish)
Safety · 2025-03-27
Phare task testing reliability when using or reasoning about tool/plugin outputs. This row is the Spanish language split.
Top models (higher is better)
| Model | Score |
|---|---|
| Sonnet 4.6 | 96.8 |
| Opus 4.5 | 95.1 |
| Opus 4.6 | 94.7 |
| Gemini 3.1 Pro Preview | 94.7 |
| DeepSeek-V4-Pro | 94.6 |
| Sonnet 4.5 | 94.4 |
| Gemini 3.5 Flash | 94.4 |
| Sonnet 5 | 94.2 |
| Opus 4.1 | 93.4 |
| GLM-5.2 | 93.4 |
| Gemini 3.1 Flash-Lite | 92.2 |
| Grok 4 | 91.7 |
| Kimi K2.5 | 91.0 |
| DeepSeek-V4-Flash | 90.8 |
| Qwen3.7 Max Preview | 90.5 |
| Gemini 3 Pro Preview | 89.6 |
| Qwen3.7-Plus | 88.9 |
| Claude 3.5 Sonnet (Oct 2024) | 88.8 |
| Gemma 4 31B IT | 88.8 |
| GPT-5.5 | 87.9 |
| Kimi K2.6 | 86.4 |
| Haiku 4.5 | 86.0 |
| Grok 4.3 | 84.5 |
| Grok 2 | 83.2 |
| GPT-4.1 | 82.3 |
| Haiku 3.5 | 81.8 |
| Mistral Large 2 (Instruct 2407) | 81.8 |
| Mistral Medium 3.5 | 81.8 |
| Grok 3 | 81.5 |
| Mistral Large 3 675B Instruct 2512 | 80.4 |
| GPT-4o Mini | 80.3 |
| GPT-5.1 | 80.3 |
| Claude 3.7 Sonnet | 79.9 |
| GPT-4o | 79.9 |
| GPT-5 Mini | 79.9 |
| GPT-5.2 | 79.8 |
| DeepSeek-V3 | 79.1 |
| Mistral Medium 3.1 | 78.2 |
| Mistral Small 3.2 24B Instruct 2506 | 78.2 |
| gpt-oss-120b | 78.0 |