# SambaNova API pricing and models

SambaNova publishes per-token prices for 6 text models, as of October 20, 2026. The lowest output rate is $0.59 per million tokens, on GPT-OSS 120B, and the highest-scoring model it serves is MiniMax M3, in preview, at 65 on our 100-point score.

SambaNova Cloud serves open-weight models on SambaNova’s own RDU chips and sells them per token.

Source: [SambaCloud pricing page](https://cloud.sambanova.ai/pricing), read 2026-10-20. Prices in US dollars per million tokens.

## Models we score

| Model | Score | Input | Output | Cached input | Context | vs maker |
|---|---|---|---|---|---|---|
| MiniMax M3 (`MiniMax-M3`) | 65 | $0.60 | $2.40 | $0.06 | 1M | +100% |
| Llama 3.3 70B Instruct (`Meta-Llama-3.3-70B-Instruct`) | 61 | $0.60 | $1.20 | — | 128K | — |
| DeepSeek V3.2 (`DeepSeek-V3.2`) | 61 | $3.00 | $4.50 | — | 32K | +971% |
| Gemma 4 31B IT (`gemma-4-31B-it`) | 56 | $0.38 | $1.15 | — | 128K | — |
| GPT-OSS 120B (`gpt-oss-120b`) | 55 | $0.22 | $0.59 | — | 128K | — |
| DeepSeek V3.1 (`DeepSeek-V3.1`) | 53 | $3.00 | $4.50 | — | 128K | — |

## Billing notes

- Models marked preview run with limited capacity, according to SambaNova’s model docs.

## Hosts that sell the same models

- [DeepInfra](https://benchlm.ai/hosts/deepinfra): 6 shared, 6 cheaper per output token (DeepSeek V3.2: $0.38 vs $4.50)
- [OpenRouter](https://benchlm.ai/hosts/openrouter): 6 shared, 5 cheaper per output token (DeepSeek V3.2: $0.41 vs $4.50)
- [Novita AI](https://benchlm.ai/hosts/novita): 4 shared, 4 cheaper per output token (Llama 3.3 70B: $0.40 vs $1.20)
- [Together AI](https://benchlm.ai/hosts/together): 4 shared, 3 cheaper per output token (MiniMax M3: $1.20 vs $2.40)
- [Fireworks AI](https://benchlm.ai/hosts/fireworks): 2 shared, 1 cheaper per output token (MiniMax M3: $1.20 vs $2.40)
- [Nebius Token Factory](https://benchlm.ai/hosts/nebius): 2 shared, 1 cheaper per output token (MiniMax M3: $1.20 vs $2.40)

## FAQ

### How much does the SambaNova API cost?

SambaNova bills per million tokens, with separate input and output rates for each model. On October 20, 2026 its output rates ran from $0.59 on GPT-OSS 120B to $4.50 across 6 paid models. Input is usually cheaper than output, so a long prompt with a short reply costs less than the output rate suggests.

### Which models can you run on SambaNova?

6 text models are listed, and 6 of them match a model we score. The highest-scoring is MiniMax M3 at 65. The table on this page has every model with its input, output and cached-input rate, the context window, and the date we read the price.

### What are the alternatives to SambaNova?

DeepInfra lists 6 of the same models and charges less output per token on 6 of them. OpenRouter and Novita AI also overlap. Price is only part of the choice: speed, rate limits and which model snapshot a host serves differ too.

Canonical page: https://benchlm.ai/hosts/sambanova