[Bug]: stream_chunk_size on bedrock/invoke models leaks into the request body and Bedrock rejects it with ValidationException
![[Bug]: stream_chunk_size on bedrock/invoke models leaks into the request body and Bedrock rejects it with ValidationException](https://www.chat-gpts.plus/wp-content/uploads/2026/06/30239-15c28b58-768x403.jpg)
用户使用 LiteLLM Proxy 路由到 Bedrock 的 invoke 模型(如 Claude Sonnet 4),在流式推理请求中设置 `stream_chunk_size` 参数( {"stream":true,"stream_chunk_size":2048} ),导致 Bedrock


![[Codex] 今天有复数个账号出现了二验手机号,询问一下有没有相同情况的?](https://www.chat-gpts.plus/wp-content/uploads/2026/06/ai_cover_3-536-768x403.jpg)




![issue: 400: [ERROR: Bad Request] when using image generation using LiteLLM openai-api](https://www.chat-gpts.plus/wp-content/uploads/2026/06/25989-7876efab-768x403.jpg)
