• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
  • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    AMD MI355x Beats B300 on AgentX Tokens per Dollar

    SemiAnalysis posted charts comparing AMD MI355x and B300 performance on AgentX.

    SE
    2 Sources, 27d ago, first seen 27d ago

    TLDR

    SemiAnalysis posted two updates highlighting a new AMD MI355x submission that delivers more total tokens per dollar of TCO than B300 at lower interactivity ranges on AgentX. The firm noted both high-throughput configurations rely on disaggregated setups. Charts attached to the posts display Kimi K3 2.8T Agentic results plotted against P90 interactivity. The account credited vLLM, AMD, and LMCache engineers for the work. No further details on test conditions or independent verification appear in the posts.

    Combined views

    42K

    2 Sources, first seen 27d ago

    Combined views

    42K

    2 Sources, first seen 27d ago

    283 likes
    283 likes
    16 comments
    29 saves
    28 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    16 comments
    29 saves
    28 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    2 Sources

    @SemiAnalysis_On total tokens per $ TCO, a new AMD MI355x submission beats B300 at lower interactivity ranges on AgentX. Both high throughput configs use disaggregated setups. Shoutout to vLLM, AMD, and LMCache engineers.

    2 Sources

    @SemiAnalysis_On total tokens per $ TCO, a new AMD MI355x submission beats B300 at lower interactivity ranges on AgentX. Both high throughput configs use disaggregated setups. Shoutout to vLLM, AMD, and LMCache engineers.