Thinking/reasoning provider options not gated by model capabilities #34323

Description

Thinking/reasoning provider options in ProviderTransform.options() are not gated by model.capabilities.reasoning. This breaks non-reasoning models in two ways:

1. Normal output from a non-thinking model gets stored as reasoning ("thinking") instead of assistant text.
2. Opting out via agent options: { thinking: { type: "disable" } } passes the field straight through to the provider, and the non-thinking model's API rejects the unknown parameter.

Seven branches in options() inject thinking/reasoning params based on provider ID or model ID matching alone. Only Google and alibaba-cn branches already gate on capabilities.reasoning.

Plugins

None

OpenCode version

1.17.11

Steps to reproduce

1. Configure a subagent using a non-reasoning model whose ID/provider matches one of the ungated branches (e.g. zai/zhipuai via openai-compatible)
2. Invoke it via the task tool — output is recorded as thinking instead of text
3. Add options: { thinking: { type: "disable" } } — provider rejects the request

Operating System

All

Terminal

All