← Timeline · Model Release

o3 and o4-mini use tools while reasoning

· significance ●●○○○

OpenAI's reasoning models can agentically search, run code and analyse images inside their chain of thought.

What happened and why it matters

Tool use inside reasoning made models markedly better at multi-step real-world tasks, though third-party evaluators also noted higher hallucination rates for o3.

Sources

Hover the chart to read values · click a marker to open its articles