Skip to content
TheChatArchiveThe Chat Archive
Perplexitysonar-reasoning· PerplexityHugging Face LMArena Search Arena 24K

What are the top ten models currently in simple-bench.com/?

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

    What are the top ten models currently in simple-bench.com/?

  2. Perplexity Assistant

    The top models currently listed on SWE-bench (a benchmarking platform for software engineering tasks) are ranked by their performance percentages as of the latest available data[3]:

    1. W&B Programmer O1 crosscheck5 (64.60%)
    2. Blackbox AI Agent (62.80%)
    3. CodeStory Midwit Agent + swe-search (62.20%)
    4. Learn-by-interact (60.20%)
    5. devlo (58.20%)
    6. Emergent E1 (v2024-12-23) (57.20%)
    7. Gru(2024-12-08) (57.00%)
    8. EPAM AI/Run Developer Agent v20241212 + Anthopic Claude 3.5 Sonnet (55.40%)
    9. ugaiforge (41.60%)
    10. nFactorial (2024-10-30) (41.60%)

    These rankings reflect performance on tasks like code generation, debugging, and software engineering workflows. The leaderboard is updated regularly, with newer models often incorporating advanced AI agents and fine-tuned language models[3].

    Note: The query references "simple-bench.com," but the provided search results only include data from SWE-bench. If the intended platform is different, additional context may be required.

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