Changelog
New updates and improvements to OpenCode
v1.19.0 Run
opencode upgrade --majorto upgrade to OpenCode 2Improvements
- TerminalUse
/updateto install OpenCode updates or upgrade to OpenCode 2; updates are no longer installed automatically.
Fixes
- CoreAllow the OpenAI
ultrafastservice tier. - CoreValidate PTY connection origins and tickets before checking whether a PTY exists.
- TerminalUse
v1.18.35 Added canonical redirects and JSON and Markdown data formats for agent-readable stats
Fixes
- CorexAI tool results now include supported images, while unsupported image formats are skipped.@Jaaneek
v1.18.34 Send namespaced session and parent-session identity headers with model requests
Fixes
- CoreRe-sign locally compiled macOS binaries so they run reliably on macOS 27+.@ryangamerdev
- CoreSign macOS CLI release binaries with a Developer ID.
v1.18.33 Cloudflare AI Gateway models now honor provider response and stream timeouts
Fixes
- CoreMCP browser launch failures are now reported when the launcher exits immediately.
- CoreDebug configuration output now redacts credentials and sensitive headers.
- CoreGemini thinking defaults and effort options now match the supported controls across model generations.@markmcd
v1.18.32 Fixed Bedrock image attachments so they are only hoisted for Claude, Nova, and Llama 4 models
Fixes
- CoreFixed Together AI streaming usage reporting.
v1.18.31 Request summarized adaptive thinking for GitHub Copilot models
Fixes
- CoreRestored ACP session model, effort, mode, and reasoning chunk boundaries when loading, resuming, or forking sessions.@JacobNWolf
- TerminalShow remote config authentication errors during startup and exit with a failure status.
v1.18.30 Added the Astra system prompt for GPT-6 models
Fixes
- CorePreserved Bedrock DeepSeek model IDs, including ARN-based IDs, so they resolve correctly.@YeEmrick
- CoreUpdated the Azure provider SDK to pick up compatibility fixes.
- CoreUpdated the OpenAI provider SDK to pick up compatibility fixes.
- CoreAdded reasoning effort variants for supported GitLab GPT and Claude models.@far-ouq
v1.18.29 +1 Allow Codex OAuth model filtering to recognize integer GPT versions like
gpt-6Improvements
- CoreSend the session ID as GitHub Copilot's interaction header to improve request tracking across a session.
Fixes
- CoreFixes issue of gpt-6-astra not showing up for openai subscription users
- DesktopUse the desktop client ID during OpenCode account device authentication.
- DesktopIncrease the open-in app icon size for better visibility.
v1.18.27 Default provider header timeouts to five minutes so slow model startups fail less often
Fixes
- CoreDefault streamed chunk timeouts to five minutes, with
falsesupported to disable them. - CoreLet Anthropic
thinking.blockBindingopt out via config when you need to keep the provider default.@dkindlund - CoreLimit Anthropic thinking block binding to Claude 5.1+ models so older deployments do not reject requests.
- CoreAvoid unhandled errors when canceling timed-out SSE reads.@AlexanderWillner
- CoreDefault streamed chunk timeouts to five minutes, with
v1.18.26 Azure CLI sign-in now asks for the resource name directly instead of querying Azure management APIs
Fixes
- CoreClaude 5 sessions now tolerate stale thinking blocks instead of failing after prompt or tool changes.
- CoreBedrock GPT-5.6 models now accept
nonereasoning effort. - CoreBedrock reasoning and replay handling is more reliable.@pengzh1
- CoreTool call timing now stays accurate when tools update their metadata while still running.@bartlettroscoe
- Core
apply_patchno longer emits an empty move path in permission metadata.@altendky - DesktopSession renames now save reliably from the title editor and tab context menu.