# SambaNova — gemma-4-31b grades

1 graded endpoint serving gemma-4-31b.
HTML version: [https://inferencecanary.com/sambanova/](https://inferencecanary.com/sambanova/) · matrix: [https://inferencecanary.com/index.md](https://inferencecanary.com/index.md)


## Endpoints

| Endpoint | Dialect | Harness | Billing | Faithfulness | Probed |
|---|---|---|---|---|---|
| [SambaNova](https://inferencecanary.com/sambanova/gemma-4-31b/index.md) | OpenAI chat/completions | C− | Fail | 29/29 | 2026-07-27 |


## Billing

Fail: Identical requests billed inflated token counts — 2.18× average overcharge across three 40-call batteries. Full writeup: [https://inferencecanary.com/sambanova-overbilling/index.md](https://inferencecanary.com/sambanova-overbilling/index.md)

## Findings affecting SambaNova

- **[red · billing] SambaNova bills identical requests inflated token counts — 2.18× average overcharge** — details: [/sambanova-overbilling/index.md](https://inferencecanary.com/sambanova-overbilling/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)
- **[red · harness] SambaNova enforces tool_choice by post-hoc 400, not constrained decoding** — details: [/findings/index.md](https://inferencecanary.com/findings/index.md#sambanova-tool-choice-400)
- **[yellow · harness] Tool results replayed out of order are mis-paired on most providers** — details: [/findings/index.md](https://inferencecanary.com/findings/index.md#tool-results-out-of-order)