Skip to content
TheChatArchiveThe Chat Archive
gpt-4o-search-preview· OpenAIHugging Face LMArena Search Arena 24K

deepseek v3 0324 gpqa benchmark? Answer in short.

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with Perplexitysonar’s answer.
  1. User

    deepseek v3 0324 gpqa benchmark? Answer in short.

  2. Assistant

    DeepSeek-V3 achieved a Pass@1 score of 59.1% on the GPQA-Diamond benchmark, surpassing GPT-4o (49.9%) and Qwen2.5 (51.1%), and closely approaching Claude-3.5's 65.0%[1]. This performance indicates DeepSeek-V3's strong capabilities in handling complex question-answering tasks.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.

Identical first prompt