Ollama: OpenAI Reasoning Integration

Ollama merged two pull requests from Parth Sareen that add support for "max" thinking values and integrate OpenAI's reasoning effort mapping. The changes include comprehensive test coverage and validation for unsupported effort values.

Duration: PT1M37S

Episode overview

This episode is a short developer briefing from Ollama.

It explains recent repository work in plain language.

  • Show: Ollama
  • Published: 2026-04-24T00:00:00Z
  • Audio duration: PT1M37S

Transcript excerpt

This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.

Good morning. This is your Ollama development briefing for April 24th, 2026.

Two pull requests were merged today, both from contributor Parth Sareen, focused on enhancing OpenAI integration capabilities.

Sareen merged "api: accept max as a think value," which adds support for models that specifically require "max" as a thinking parameter. The changes span six files including API types, command handling, and server routes, with 51 new test cases added to ensure compatibility across the OpenAI integration layer.

The second merge, "openai: map responses reasoning effort to think," builds on this foundation by mapping OpenAI's Responses API reasoning effort levels directly into Ollama's chat request thinking parameter. This implementation preserves explicit "max" reasoning for renderers that need it while returning validation…

Both changes modify core API types and OpenAI integration code, with the server routes updated to handle the new thinking value options. The modifications add 180 lines of new code with no deletions, indicating these are purely additive features.

What's next: These changes position Ollama to better integrate with OpenAI's reasoning models and provide more granular…

Nearby episodes from Ollama

  1. Multi-Sequence Batching and New Model Support
  2. Tokenizer Bug Fix for BPE Processing
  3. Weekly Recap - MLX Performance & Launch Integrations
  4. MLX Sampling Performance Enhancement
  5. Launch System Improvements and Integration Fixes
  6. Launch System Overhaul and Documentation Updates
  7. MLX Performance Boost and Model Updates
  8. New CLI Integration and Performance Improvements