Unverified60% confidenceFactTime unknown
在Rios Comp Agent Swarm测试中，K2.6拿下86.3分，GPT-5.4只有78.4分，差距接近8个百分点
2
Sources
60%
Confidence
Medium-term (~90 days)
Relevance
5/31/2026
First Seen
Valid until: 8/29/2026
Sources
Kimi K2.6开源实测：300个Agent协同的调度能力到底多强
bilibiliAi学习的老章
Related Entities
Related Claims
Verified在Terminal Bench 2.0测试中，K2.6打出66.7分，比GPT-5.4和Claude Opus 4.6都略高71% similarUnverified在综合性能排名中，Kimi K2.5得分64%，排名第五，前四名分别是Gemini 3 Pro（100%）、Claude Opus 4.5 Max（74%+）、GLM 4.7（65%）和GPT-5.2X（65%）68% similarUnverified在Artificial Analysis第三方评测的编码指数（推理模式）上，GPT-5 Mini High的得分比Haiku 4.5高出9个点66% similarUnverified在Terminal Bench 2.1基准测试上，GPT-5.6 Sol的Ultra模式接近92%，超越竞品的88%表现66% similarPartially VerifiedClaude Opus 4.8在SWE-bench基准测试中得分69.2%，GPT 5.5为58.6%，Google Gemini为54.2%65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/4033API
curl https://kongchang.com/api/v1/knowledge/claims/4033MCP
get_claim(id=4033)