# AWS — gemma-4-31b grades

2 graded endpoints serving gemma-4-31b.
HTML version: [https://inferencecanary.com/aws/](https://inferencecanary.com/aws/) · matrix: [https://inferencecanary.com/index.md](https://inferencecanary.com/index.md)

A vendor is graded once per API it speaks, because a shim and a native protocol routinely
behave differently — AWS speaks 2.


## Endpoints

| Endpoint | Dialect | Harness | Billing | Faithfulness | Probed |
|---|---|---|---|---|---|
| [AWS Bedrock (Chat Completions)](https://inferencecanary.com/aws/gemma-4-31b/index.md) | OpenAI chat/completions on Bedrock Mantle | D | OK | 28/29 | 2026-07-27 |
| [AWS Bedrock (Responses)](https://inferencecanary.com/aws-responses/gemma-4-31b/index.md) | OpenAI Responses on Bedrock Mantle | D− | OK | 25/29 | 2026-07-27 |


## Billing

Footnote (customer-favoring, not flagged): Reports reasoning_tokens: 0 beside delivered reasoning — under-counts in the customer's favor.

## Findings affecting AWS

- **[red · harness] Four providers drop replayed reasoning: the model never sees its own prior thinking** — details: [/replayed-reasoning-dropped/index.md](https://inferencecanary.com/replayed-reasoning-dropped/index.md)
- **[red · harness] Images in tool results: rejected, 500'd, or silently blinded on 7 of 11 providers** — details: [/findings/index.md](https://inferencecanary.com/findings/index.md#tool-result-images)
- **[yellow · harness] AWS Bedrock never emits parallel tool calls** — details: [/findings/index.md](https://inferencecanary.com/findings/index.md#bedrock-single-tool-call)
- **[yellow · harness] AWS Bedrock un-escapes tool-call arguments in transit** — details: [/findings/index.md](https://inferencecanary.com/findings/index.md#bedrock-argument-corruption)
- **[yellow · harness] AWS Bedrock (Responses) rejects temperature** — details: [/findings/index.md](https://inferencecanary.com/findings/index.md#aws-responses-temperature)