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

[models-bot]

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

---

Note

Low Risk
Adds static model metadata only; no runtime, auth, or application logic changes.

Overview
Adds a new AWS Bedrock provider definition for anthropic.claude-opus-5, enabling routing, pricing, and capability metadata for the model.

The entry marks the model active and serverless, with chat mode, 1M context, 128k max output, text/image input, and extended thinking (thinking: true). It documents tool use, prompt caching, and related Anthropic features, plus tunable max_tokens and reasoning_effort (default high). Token costs (including batch and cache pricing) are specified for ap-northeast-2, with links to AWS and Anthropic docs in sources.

^{Reviewed by Cursor Bugbot for commit dd4b001. 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 dd4b001. 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.