PulseAugur
EN
LIVE 16:44:23

Moonshot releases Kimi K3, a 2.8T parameter multimodal model with 1M context

Moonshot has released Kimi K3, a new 2.8 trillion parameter multimodal model featuring a 1 million token context window and native vision capabilities. The model demonstrates impressive speed, achieving 460 tokens per second upon release. Kimi K3 is built on a mixture-of-experts architecture and has been made available through Modal's platform, offering both token-based pricing and dedicated capacity options. AI

IMPACT Sets a new benchmark for open-source multimodal models with large context windows, potentially accelerating agentic applications.

RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on Modal blog →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Moonshot releases Kimi K3, a 2.8T parameter multimodal model with 1M context

COVERAGE [1]

  1. Modal blog TIER_1 English(EN) ·

    Kimi K3 by Moonshot now available on Modal

    Kimi K3, a 2.8 trillion parameter multimodal model by Moonshot, along with a custom-trained DFlash speculator, is now available on Modal.