Chinese AI lab Moonshot released Kimi K3 on July 16, a 2.8-trillion-parameter open-weight multimodal reasoning model — the largest open-weight model shipped to date. Two variants launched first inside the Kimi app and Kimi Code: K3 Max for chat and agent tasks, and K3 Swarm Max for large-scale parallel processing.
The model ships with a 1-million-token context window, an always-on "thinking mode," and aggressive pricing at $3 per million input tokens and $15 per million output tokens. On independent benchmarks it lands fourth among all frontier models, trailing only Claude Fable 5 and GPT-5.6 Sol while edging past Claude Opus 4.8. Full open weights are due by July 27.
The release lands as Moonshot raises new funding at a $31.5 billion valuation, up from $20 billion in May, and marks the latest sign that Chinese labs are closing the gap with US frontier AI amid a broader industry push toward open-weight alternatives.