Meet Kimi K2.6 Agent Swarm 👋 Highlights: 🔹 Swarms, elevated - 300 parallel sub-agents × 4,000 steps per run (up from 100 / 1,500 in K2.5). 🔹 Outputs are real files, not chat - one run delivers 100+ files, 100,000-word literature reviews, or 20,000-row datasets. 🔹Heterogeneous skills - search, analysis, coding, long-form writing, and visual generation all running in parallel 🔗Try it at: https://t.co/2Tu8McUaUa
Moonshot AI Scales Kimi K2.6 Agent Swarm to 300 Parallel Subagents
· Updated
Moonshot AI released Kimi K2.6 Agent Swarm, a system that coordinates up to 300 parallel sub-agents across 4,000 steps to complete complex tasks. The update shifts from conversational responses to production-ready file bundles, including 100-page reports and massive datasets.
- Parallel sub-agents
- 300
- Coordinated steps
- 4,000
- Max file output
- 100+ files
- Dataset capacity
- 20,000 rows
- Literature review length
- 100,000 words
- Previous sub-agent limit
- 100
- Previous step limit
- 1,500
This update addresses the complexity ceiling where agents lose coherence during long tasks. By using heterogeneous skills in parallel, the system moves beyond chat. It follows the long-horizon coding capabilities of the base model and extends the optimized prefill processing required for massive context windows.
You can now generate asset bundles from one prompt, such as a 100-page PDF report and a 20,000-row dataset. This builds on distributed inference architectures designed to handle the massive compute required for long-running tasks. You can access the swarm mode through the official Kimi interface.
Still wondering? A few quick answers below.
Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →





