SambaNova API Pricing (2026)
SambaNova Cloud delivers ultra-fast AI inference on custom-built DataScale chips, achieving some of the highest tokens-per-second throughput in the industry. Specializes in serving open-source models like Llama and DeepSeek at competitive prices with enterprise-grade reliability. Founded in 2017 and headquartered in Palo Alto, CA, SambaNova offers 5 active models through their API. All pricing data below is sourced directly from the SambaNova pricing page and is updated regularly.
Prices verified Mar 11, 2026
SambaNova Model Pricing — All Models
All prices are in USD per 1 million tokens ($/M tokens). Input tokens are the text you send to the model; output tokens are the generated response.
| Model | Tier | Input | Output | Cached Input | Batch Input | Batch Output |
|---|---|---|---|---|---|---|
Llama 3.1 405B (SambaNova) Meta's largest Llama 3.1 model hosted on SambaNova Cloud's custom DataScale chips. Delivers top-tier open-source model performance with ultra-high throughput, making it suitable for demanding enterprise workloads and complex reasoning tasks. | Premium | $5.00/M | $10.00/M | — | — | — |
Llama 3.1 70B (SambaNova) Meta's Llama 3.1 70B model on SambaNova Cloud. Combines strong instruction-following and reasoning capabilities with cost-effective pricing and SambaNova's high-speed inference for production-ready applications. | Mid | $0.60/M | $1.20/M | — | — | — |
Llama 3.1 8B (SambaNova) Meta's compact Llama 3.1 8B model on SambaNova Cloud. Provides lightning-fast inference at budget-friendly prices, ideal for high-volume applications like classification, extraction, and simple conversational tasks. | Budget | $0.10/M | $0.20/M | — | — | — |
DeepSeek V3 (SambaNova) DeepSeek V3 served on SambaNova Cloud. A powerful mixture-of-experts model from DeepSeek with strong coding and reasoning capabilities, enhanced by SambaNova's high-throughput infrastructure for enterprise-grade deployment. | Mid | $1.00/M | $1.50/M | — | — | — |
GPT-OSS 120B (SambaNova) OpenAI's open-source GPT model variant served on SambaNova Cloud. A 120B parameter model optimized for instruction following, available through SambaNova's custom chip infrastructure for high-throughput enterprise inference. | Mid | $0.22/M | $0.59/M | — | — | — |
Prices last verified: 2026-03-11. Always confirm on the official pricing page before production use.
SambaNova Model Capabilities
Key technical capabilities across all SambaNova models. Use this table to identify which models support the features your application requires.
| Model | Context | Max Output | Vision | Functions | JSON Mode | Caching | Batch API | Reasoning |
|---|---|---|---|---|---|---|---|---|
| Llama 3.1 405B (SambaNova) | 128,000 | 8,192 | — | ✓ | ✓ | — | — | — |
| Llama 3.1 70B (SambaNova) | 128,000 | 8,192 | — | ✓ | ✓ | — | — | — |
| Llama 3.1 8B (SambaNova) | 128,000 | 8,192 | — | ✓ | ✓ | — | — | — |
| DeepSeek V3 (SambaNova) | 128,000 | 8,192 | — | ✓ | ✓ | — | — | — |
| GPT-OSS 120B (SambaNova) | 128,000 | 8,192 | — | ✓ | ✓ | — | — | — |
When to Choose SambaNova
Cost-sensitive workloads
Use Llama 3.1 8B (SambaNova) — SambaNova's most affordable model at $0.10/M input / $0.20/M output. Estimated cost at 100K requests/month: $20.00.
Complex or high-accuracy tasks
Use Llama 3.1 405B (SambaNova) — SambaNova's most capable model (premium tier) with a 128,000-token context window. Priced at $5.00/M input / $10.00/M output.
High-volume pipelines
SambaNova models are well-suited for high-throughput workloads. Consider batch API (where supported) for an additional 50% cost reduction on asynchronous pipelines. Check the pricing table above for batch pricing availability.
Compare SambaNova Models vs Competitors
See detailed cost comparisons between SambaNova models and models from other providers, including side-by-side pricing tables and monthly cost projections.
Frequently Asked Questions: SambaNova API Pricing
What is the cheapest SambaNova model?
The cheapest SambaNova model by input token price is Llama 3.1 8B (SambaNova), priced at $0.10/M for input tokens and $0.20/M for output tokens. It is best suited for high-volume, cost-sensitive workloads where speed and price matter most.
What is the most capable SambaNova model?
Llama 3.1 405B (SambaNova) is SambaNova's most capable model, classified as a premium-tier model. It supports a context window of 128,000 tokens and is priced at $5.00/M input / $10.00/M output per million tokens.
How does SambaNova API pricing work?
SambaNova charges per token consumed — separately for input tokens (the text you send) and output tokens (the text the model generates). Prices are listed in USD per 1 million tokens ($/M). SambaNova was founded in 2017 and is headquartered in Palo Alto, CA. All pricing shown here is sourced from the official SambaNova pricing page.
Does SambaNova offer batch API pricing?
SambaNova's current model lineup does not include batch API pricing at this time. For cost reduction on high-volume workloads, consider prompt caching (where supported) or running multiple concurrent requests with standard pricing.
Do SambaNova models support vision / image inputs?
None of the current SambaNova models in our dataset support vision input. SambaNova models are text-focused. For multimodal workloads requiring image understanding, consider comparing models from other providers using our comparison tool.
SambaNova pricing data last verified: 2026-03-11. Prices may change — verify on the official SambaNova pricing page before production use.