GPT-6 Sol Ultra Subagents Use Up Limits Fast

Melvin Vivas · X post · 2026-09-24 · Open on X

Topics: AI Agents, Tool Use & MCP, LLMOps, Deployment & Monitoring · Level: intermediate

Summary

An observation that GPT-6 Sol Ultra spawns subagents at the same Ultra reasoning level, which uses up usage limits quickly. A cost warning for multi-agent coding workflows.

Key points

• GPT-6 Sol Ultra often spawns Sol subagents at the same Ultra reasoning level.
• High-reasoning subagents multiply token use and use up plan limits fast.
• Keep an eye on subagent reasoning levels to control cost.

Resources mentioned

• Luna · tool · openai.com · paid
  The other model/agent the classifier routes tasks to; the post doesn't describe it further.
  Also in: GPT-6.1 Sol May Beat Luna for Subagents (Melvin Vivas on X · notes), Picking models for orchestrator and subagent roles in Codex (Melvin Vivas on X · notes), GPT-6 Sol vs Opus 5.5 in a Livestream Comparison (Melvin Vivas on X · notes), DigitalOcean Serverless Inference Now Serves OpenAI GPT-6 Models (Melvin Vivas on X · notes) and 14 more