Cherry-pick: Add GLM-4.7 model support to Z.ai provider

[kiloconnect]

This PR cherry-picks the changes from RooCodeInc/Roo-Code#10271

Changes

• Adds GLM-4.7 model support to Z.ai provider (both international and mainland China configurations)
• Adds defaultToolProtocol: "native" to all Z.ai models for improved tool calling
• Includes changeset documentation

Original Author

Original commit by: Roo Code roomote@roocode.com

Testing

• Verified that the model configurations are properly added
• Ensured original author attribution is preserved in git history

[changeset-bot]

🦋 Changeset detected

Latest commit: 8121f0a

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
kilo-code	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[kiloconnect]

✅ No Issues Found

2 files reviewed | Confidence: 95% | Recommendation: Merge

Review Details

Files:

• .changeset/add-glm-4-7-model.md - Changeset for patch version bump
• packages/types/src/providers/zai.ts - Model configuration additions

Changes Verified:

• ✅ GLM-4.7 model added to both international and mainland configurations
• ✅ defaultToolProtocol: "native" added to all models with supportsNativeTools: true
• ✅ Model structure follows existing patterns (contextWindow, pricing, etc.)
• ✅ Type compatibility verified against ModelInfo schema

Checked: Security, bugs, type safety, consistency with existing patterns

[kiloconnect]

✅ No Issues Found

2 files reviewed | Confidence: 95% | Recommendation: Merge

Review Details

Files:

• .changeset/add-glm-4-7-model.md - Standard changeset documentation
• packages/types/src/providers/zai.ts - Model configuration additions

Changes Verified:

• ✅ New GLM-4.7 model added for both international and mainland China APIs
• ✅ defaultToolProtocol: "native" added consistently to all Z.ai models with supportsNativeTools: true
• ✅ Model configuration follows established patterns (pricing, context windows, capabilities)
• ✅ Type safety maintained - values conform to ModelInfo schema

Checked: Security, type safety, pattern consistency, breaking changes

[Bigpet]

Should it be added here too?

kilocode/src/api/transform/model-params.ts

Lines 195 to 203 in 6a4b8fd

	// kilocode_change start
	function shouldDisableReasoning(modelId: string, reasoningEffort: ReasoningEffortExtended | undefined) {
	const supportsReasoningToggle =
	modelId.startsWith("deepseek/deepseek-v3.1") ||
	modelId.startsWith("deepseek/deepseek-chat-v3.1") ||
	modelId.startsWith("x-ai/grok-4-fast") ||
	modelId.startsWith("z-ai/glm-4.6")
	return supportsReasoningToggle && reasoningEffort === "minimal"

[mcowger]

Should it be added here too?

kilocode/src/api/transform/model-params.ts

Lines 195 to 203 in 6a4b8fd

	// kilocode_change start
	function shouldDisableReasoning(modelId: string, reasoningEffort: ReasoningEffortExtended | undefined) {
	const supportsReasoningToggle =
	modelId.startsWith("deepseek/deepseek-v3.1") ||
	modelId.startsWith("deepseek/deepseek-chat-v3.1") ||
	modelId.startsWith("x-ai/grok-4-fast") ||
	modelId.startsWith("z-ai/glm-4.6")
	return supportsReasoningToggle && reasoningEffort === "minimal"

Don't think so - reasoning seems to be fine on 4.7