Cost Comparison of Models in Agentic Workflows
rohanpaul_ai · x · 2026-07-11
This post primarily compares the performance of several models in agentic workflows, meaning specialized tasks requiring long-term planning, tool calling, checking, and retries.
Key conclusions:
- GPT-5.6 Sol is considered the best fit for such workflows
- It achieves roughly 53%, at a cost of about $500–$1,000
- Claude Opus 4.8 achieves a lower score at a cost of about $1,400–$4,000
- Claude Fable 5 costs about $2,400, scoring around 40%
Quoted official OpenAI information also mentions:
- ChatGPT Work has started rolling out to Pro / Enterprise / Edu users
- It will roll out to Plus / Business in the coming days
- On desktop, Chat / Work / Codex are available across all tiers, including Free
- Windows and Mac versions are globally available for download, and the Codex app will be updated to the new ChatGPT desktop app
Related event: Significant Cost Disparities Among Models in Long-Context Agent Tasks(3 posts)→
More from coding & agent
- A roundup of AI agents and MCP resources, including how to evaluate agents — _jaydeepkarale · 2026-07-21
- Anthropic shares a masterclass on how it builds AI agents — _jaydeepkarale · 2026-07-21
- Anthropic masterclass spotlights how to build and observe AI agents — _jaydeepkarale · 2026-07-21
- A beginner guide to AI agents points readers to a Stanford webinar — _jaydeepkarale · 2026-07-21
- A full course shows how to build and deploy an AI agent with OpenAI and LangChain — _jaydeepkarale · 2026-07-21
- A practical guide on how to evaluate AI agents — _jaydeepkarale · 2026-07-21