Open-ended Question Answering on HLE
20.5AccuracyAnthropic
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| AnthropicProvider=Anthropic, Sample Size (N)=14862026.05 | 20.5 | 0.4563 | 0.4574 | 0.4403 | |
| GeminiProvider=Gemini, Sample Size (N)=21112026.05 | 15.1 | 0.7917 | 0.5362 | 0.5227 | |
| OpenAIProvider=OpenAI, Sample Size (N)=21282026.05 | 10.1 | 0.6804 | 0.4696 | 0.4522 | |
| xAIProvider=xAI, Sample Size (N)=20052026.05 | 6.9 | 0.7317 | 0.5984 | 0.584 | |
| MistralProvider=Mistral, Sample Size (N)=21282026.05 | 5.3 | 0.8143 | 0.6868 | 0.6749 |