The opencode under WSL outputs only one word per line during thinking.

[MSoberM]

Description

The opencode under WSL outputs only one word per line during thinking.

Plugins

No response

OpenCode version

1.3.9

Steps to reproduce

No response

Screenshot and/or share link

No response

Operating System

No response

Terminal

No response

[HRronaldo]

same
win11 desktop & win11 terminal

opencode v1.3.13

[Augustine-C]

Encountered similar problems here using @ai-sdk/openai-compatible. With custom model provider that returns thinking content in choices.delta.reasoning_content
linux & win11
OpenCode v1.3.13

[sghng]

Similar issue encountered on macOS with GLM-5 from OpenCode Zen. This might be an upstream problem, not a OpenCode problem?

[Augustine-C]

Similar issue encountered on macOS with GLM-5 from OpenCode Zen. This might be an upstream problem, not a OpenCode problem?

I am using self-deployed models and got the same error

[roelwuytens]

Same on Linux with MiniMax M2.5 (OpenCode Zen) with OpenCode version 1.3.13

[pava0]

Same with MiniMax M2, Cursor and tmux on ssh