Discussing Sub-Agents and Reasoning Mode Costs
reach_vb · x · 2026-07-12
The post discusses practical behaviors within a coding/agent workflow:
- It is pointed out that GPT 5.6 Sol only uses subagents in Ultra mode, unless explicitly prompted in other modes.
- It suggests testing medium / high / xhigh based on different experiments and workflows to observe the required reasoning effort.
- Another reply mentions that excessive triggering of subagents in certain "code-fixing" loops can incur unnecessary costs. It advises considering a switch back to a token-based approach and trying to write guidance into skill or subagent configurations to limit reasoning intensity.
Related event: GPT-5.6 Subagent Mechanism Sparks Cost and Over-triggering Controversy(5 posts)→
More from coding & agent
- Alex Townsend posts 200 open problems in numerical linear algebra for humans and AI agents — IgorCarron · 2026-09-11
- Kimi K2.8 Preview rolls out: near-K3 coding performance, 1M context for all tiers — teortaxesTex · 2026-09-11
- Looking for a classifier of software engineering task shapes to pick models per task — StewartalsopIII · 2026-09-11
- Steal this idea: prompt-to-hardware where agents assemble custom devices — paraschopra · 2026-09-11
- Model Is the Least Interesting Part: A Guide to Six Core AI Architectures from RAG to Multi-Agent — goyalshaliniuk · 2026-09-11
- Non-coder builds layered memory architecture: 20k tokens tracks a year of agent conversations — matteoianni · 2026-09-11