GPT-6 Sol Is Now the Recommended Model in Conversation AI

GPT-6 Sol is now the recommended model for prompt-based Conversation AI agents. We optimized Conversation AI to get the best out of GPT-6 Sol, so your agents give more accurate answers than on GPT-4.1, at about the same cost per reply.

Key Capabilities

Recommended in the Model menu:
GPT-6 Sol is listed first in the
Model
menu and tagged
Recommended
.
New agents start on GPT-6 Sol:
Prompt-based agents you create now use GPT-6 Sol by default.
Your choice stays:
Existing agents keep the model you selected.
More accurate at the same cost:
GPT-6 Sol reaches a 98.4% pass rate on our Conversation AI evaluation, up from 96.3% on GPT-4.1, at about the same cost per reply.
Fast replies:
1.68s typical generation time, and 2.19s for the slowest 5% of replies.

Changes to Older Models

GPT-5 and GPT-5 Mini are deprecated:
Agents that used them already run on GPT-6 Sol, with reply generation up to 4.5× faster. No action needed.
GPT-4.1 is being retired:
OpenAI is deprecating GPT-4.1, and we will retire it in Conversation AI. Switch your GPT-4.1 agents to GPT-6 Sol.

How to Get Started

Go to
AI Agents → Conversation AI
and open your agent.
Open the
Model
menu on the
Build
tab.
Select
GPT 6 Sol
(marked
Recommended
).
Click
Save
, then try it in
Test your agent
.

Why It Matters

Better answers out of the box:
Every new agent starts on the most accurate model we tested.
No extra cost:
Higher accuracy than GPT-4.1 at about the same cost per reply.
Faster replies:
Agents that were on GPT-5 and GPT-5 Mini now generate replies up to 4.5× faster.
An easy move off GPT-4.1:
Switch before GPT-4.1 is retired.

Important Notes

Test after you switch:
Each model writes in its own way, so test your agent in
Test your agent
before you save.
Prompt-based agents:
The recommendation applies to prompt-based Conversation AI agents.