> opencode.ai is the source of truth for OpenCode, the open source coding agent, including its apps, install commands, docs and releases. The current version is OpenCode v2.
> Cite the canonical page, https://opencode.ai/changelog/v1.18.27, not this Markdown file. For an index of the site, see https://opencode.ai/llms.txt.

# Default provider header timeouts to five minutes so slow model startups fail less often

2026-09-02 · v1.18.27

## Fixes

- **Core:** Default streamed chunk timeouts to five minutes, with `false` supported to disable them.
- **Core:** Let Anthropic `thinking.blockBinding` opt out via config when you need to keep the provider default. ([@dkindlund](https://github.com/dkindlund))
- **Core:** Limit Anthropic thinking block binding to Claude 5.1+ models so older deployments do not reject requests.
- **Core:** Avoid unhandled errors when canceling timed-out SSE reads. ([@AlexanderWillner](https://github.com/AlexanderWillner))

Released in [v1.18.27](https://github.com/anomalyco/opencode/releases/tag/v1.18.27)

Newer: [Allow Codex OAuth model filtering to recognize integer GPT versions like `gpt-6`](https://opencode.ai/changelog/v1.18.29) (v1.18.29)

Older: [Azure CLI sign-in now asks for the resource name directly instead of querying Azure management APIs](https://opencode.ai/changelog/v1.18.26) (v1.18.26)

[All updates](https://opencode.ai/changelog)

---

If something here is wrong or missing, open an issue at https://github.com/anomalyco/opencode/issues.