Ollama: Responses Compatibility and Clearer Errors
Most activity centers on making the Responses API and ChatGPT gateway tolerate real-world history and report auth and payment failures clearly. Supporting fixes address tool-call parsing, model runtimes, and proxy correctness.
Duration: PT2M34S
Episode overview
This episode is a short developer briefing from Ollama.
It explains recent repository work in plain language.
- Show: Ollama
- Published: 2026-10-15T13:01:05Z
- Audio duration: PT2M34S
Transcript excerpt
This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.
Good morning, it's October 15th, 2026 — this is the Ollama daily briefing.
The throughline is compatibility hardening around the Responses API and ChatGPT handoffs, with less strict history validation and clearer authentication and payment failures.
First, history tolerance. Continuing a Responses conversation could fail before inference if it carried custom tool calls, another provider's compaction record, or leftover reasoning effort from a thinking model. Merged PR 18911 and open PRs 19003 and 19010 accept those items, preserve call IDs and message order, or…
Second, error transparency. Merged PRs 19004 and 19007 stop rewriting authentication and subscription failures in the local server and gateway, returning a clear sign-in message with structured metadata intact. The desktop app also distinguishes payment errors instead of treating every 402 as exhausted credit, per…
Finally, tool reliability and runtimes. Open PR 19028 fixes lost tool-call opening tags split across streaming chunks in three parsers, addressing issue 18681. Related work preserves Qwen tool calls when structured output is requested, and proposes a runtime update to support MiniCPM version 4.7 loading, while the…
W…
Nearby episodes from Ollama
- Proxy Transfers and API Resilience
- Model Compatibility Catch-Up and Windows Reliability
- Decision Models Go Usable, MLX Speeds Up on CUDA
- Decision Scoring Matures as Inference Hardens
- Native Decision Models and Precision Fixes
- Weekly Recap - Decision Models Go Native, Embeddings Go Multimodal
- GPU Discovery Fallback and Docs Cleanup
- OpenAI Parity and Reasoning Fixes