Backport: add caching provider option to GatewayProviderOptions

[n1ckoates]

Background

The caching provider option for the AI Gateway was previously untyped despite being in AI Gateway docs.

#15768 added it, but only to v7

Summary

Backported that change by updating the gatewayProviderOptions zod schema in v6.

Manual Verification

This is properly typed in v6 now with intellisense working and no errors:

  const result = streamText({
    model: 'anthropic/claude-sonnet-4.6',
    system:
      'You are a helpful assistant with access to a large knowledge base...',
    prompt: 'What is the capital of France?',
    providerOptions: {
        gateway: {
            caching: 'auto',
        } satisfies GatewayProviderOptions,
    },
  });

Checklist

• All commits are signed (PRs with unsigned commits cannot be merged)
• Tests have been added / updated (for bug fixes / features)
• Documentation has been added / updated (for bug fixes / features)
• A patch changeset for relevant packages has been added (for bug fixes / features - run pnpm changeset in the project root)
• I have reviewed this pull request (self-review)

Related Issues

Fixed #14595

Closes #14609
Closes #14632
Closes #14664
Closes #14886
Closes #15298
Closes #15329
Closes #15458
Closes #15500
Closes #15857

[github-actions]

🚀 Published in:

Package	Version
ai	6.0.216 github npm
@ai-sdk/angular	2.0.217 github npm
@ai-sdk/gateway	3.0.140 github npm
@ai-sdk/langchain	2.0.224 github npm
@ai-sdk/llamaindex	2.0.216 github npm
@ai-sdk/moonshotai	2.0.30 github npm
@ai-sdk/react	3.0.218 github npm
@ai-sdk/rsc	2.0.216 github npm
@ai-sdk/svelte	4.0.216 github npm
@ai-sdk/vue	3.0.216 github npm