Expired70% confidenceFactTime unknown
Kimi K2.5 topped the open-source leaderboard in the AI coding and multimodal understanding tracks.
1
Sources
70%
Confidence
Medium-term (~90 days)
Relevance
5/31/2026
First Seen
Valid until: 8/29/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedKimi K2.5 claims the top spot among open-source models on AI programming and multimodal understanding benchmarks84% similarUnverifiedKimi K2.5 offers multiple online usage modes including Instant (fast response), Thinking (complex reasoning), and Agent (intelligent agent).78% similarUnverifiedIn real-world testing, Kimi K2.5 automatically spun up multiple agents including a requirements research Agent, a competitive analysis Agent, a technical solution Agent, a differentiation strategist, and a UI designer to collaboratively complete the task.78% similarUnverifiedKimi K2.5 can automatically break down tasks, assign roles, and progressively advance based on a single-sentence requirement, ultimately outputting a runnable website.76% similarPartially VerifiedKimi K2.5是大规模混合专家（MoE）架构的语言模型75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/4319API
curl https://kongchang.com/api/v1/knowledge/claims/4319MCP
get_claim(id=4319)