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

lmsys vs lmarena

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

    lmsys vs lmarena

  2. Perplexity Assistant

    LMSYS and LMArena are closely related but distinct entities in the AI research ecosystem. Here's a breakdown of their relationship and roles:

    Organizational Structure

    • LMSYS is a research collective at UC Berkeley, originally a student club that developed projects like Vicuna, Chatbot Arena, and RouteLLM[1][4]. It serves as an incubator for open-source AI initiatives.
    • LMArena (formerly Chatbot Arena) is a specific project developed by LMSYS members, now spun out as an independent platform while maintaining partnership with LMSYS[3]. It focuses on live, community-driven LLM evaluation through pairwise comparisons.

    Key Differences

    AspectLMSYSLMArena (Chatbot Arena)
    ScopeBroad research collectiveSpecialized evaluation platform
    ProjectsVicuna, SGLang, RouteLLM, etc.Benchmarking LLMs via Elo ratings
    GovernanceAcademic research groupIndependent entity with own site
    FocusModel development & toolsModel evaluation infrastructure

    Relationship Timeline

    1. Origins: Chatbot Arena was launched in May 2023 as an LMSYS project to address static benchmark limitations[2][4].
    2. Growth: It became a de facto standard for LLM evaluation, attracting over 1 million participants and evaluating 90+ models[2].
    3. Spinout: In September 2024, LMArena graduated to its own platform (lmarena.ai) to focus on expanding evaluation capabilities while remaining a LMSYS partner[3].

    Shared Vision

    Both aim to advance open AI systems through:

    • Transparency: Open-source infrastructure and data sharing[2][4]
    • Community Engagement: Crowdsourced evaluations and collaborative development[1][2]
    • Innovation: Moving beyond text to multimodal and specialized evaluations[1][3]

    The distinction lies in their operational focus - LMSYS as a research incubator and LMArena as its specialized evaluation arm now operating independently[3].

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