feat(aws-bedrock/in.anthropic.claude-sonnet-5): add new models [bot]

[models-bot]

Auto-generated by model-addition-agent for aws-bedrock/in.anthropic.claude-sonnet-5.

---

Note

Low Risk
Metadata-only model registry addition with no runtime or authentication changes.

Overview
Adds a new AWS Bedrock provider catalog file for in.anthropic.claude-sonnet-5, enabling routing and cost tracking for the India inference profile of Claude Sonnet 5.

The entry marks the model active for chat with 1M context and 128k max output, text/image input, and extended thinking enabled. Pricing is defined for ap-south-1 and ap-south-2, including batch and prompt-cache token rates, and capability flags cover tools, system messages, and cache control.

^{Reviewed by Cursor Bugbot for commit 2c267a8. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 2c267a8. Configure here.}

[Archit-Kumar-16]

Already handled by the gateway without the flag. normalizeSystemRoleMessages ORs the feature against claudeVersionKeepsMidConversationSystem() (src/utils/modelParamUtils.ts:476-480), whose version table is:

MID_CONVERSATION_SYSTEM_FROM = { fable: [5,0], mythos: [5,0], opus: [4,8], sonnet: [5,0] }

Opus 5 and Sonnet 5 both clear that threshold, so mid-conversation role: system messages are preserved for these ids as written. Adding the explicit feature would be harmless but changes no behaviour.