What’s New in Ollama 0.40.14: Claude Code Output Cap Fix and Saner Sampler Defaults

Ollama 0.40.14 is a fix release.

• ollama launch claude now applies the model’s output token limit for Claude Code, so cloud models no longer hit the 128k default cap
• Cloud web search responses on /api/chat now decode correctly, including compressed ones
• OpenAI-compatible endpoints no longer override a model’s packaged sampler defaults. Clients that send no temperature/top_p get the model’s own values instead of 1.0
• /v1/responses keeps readable input content, including earlier assistant refusals
• The app can retry ChatGPT disconnect failures, and ollama launch codex checks ChatGPT routing first
• Faster top-k sampling on MLX for big vocabularies

My take

The sampler fix is the quiet important one. If local model output through the OpenAI API felt off, this might be why.

Review before you install: read the release notes yourself, back up your config, and try new versions somewhere you can afford to break first.

God Bless, and May the Source Be with You!

References & Further Reading

• Ollama v0.40.14 release (GitHub)
• What’s New in Ollama 0.40.13 (EtherNinja)