STORY · MODELLER_
Moonshot Kimi K2.5 – Chinese open-weights model beats Claude Sonnet at half the cost
Moonshot AI launched Kimi K2.5, a 32 billion parameter open-weights model with native image and video understanding, trained on 15 trillion multimodal tokens. The model offers Agent Swarm functionality to coordinate up to 100 parallel agents and claims SOTA results on multiple benchmarks with very aggressive API pricing.
WHY IT MATTERS
This represents a significant step in the arms race between Chinese and Western AI labs – an open-weights model claiming better performance than Claude Sonnet at half the cost could reshape market dynamics for API users and unlock new use cases for automation with agent swarms.
SOURCES
MACHINE-GENERATED SUMMARY This summary is written by machine from the sources below. We sort and explain — but we are a way into the field, not the final word. Check the source when something matters to you.