Groq vs SambaNova: API cost and speed
SambaNova delivers GPT-OSS 120B at 708.3 tokens/s against Groq’s 469.9 tokens/s; Groq trades that output speed for a smaller $0.0021 request example, versus $0.00279.123
| Dimension | Winner | Margin | Evidence |
|---|---|---|---|
| Estimated request cost | Groq | $0.00069 | Groq $0.0021 / SambaNova $0.00279 |
| Output speed (tokens/s) | SambaNova | 238.4 tokens/s | Groq 469.9 / SambaNova 708.3 tokens/s |
| First-chunk median (s) | Groq | 0.57 s | Groq 0.66 s / SambaNova 1.23 s |
| Endpoint Accuracy Index v1.0 | SambaNova | 12 pp | Groq 86% / SambaNova 98% |
Performance
Prices and request cost
| Provider | Input / 1M tokens | Output / 1M tokens |
|---|---|---|
| Groq2 | $0.15 | $0.6 |
| SambaNova3 | $0.22 | $0.59 |
| Product documentation | Groq API documentationSambaNova Cloud documentation | |
| API documentation | Novita AI · Chat API | |
Capabilities and serving conditions
| Dimension | Groq | SambaNova |
|---|---|---|
| Compared model | GPT-OSS 120B · | GPT-OSS 120B |
| API model ID | openai/gpt-oss-120b2 | gpt-oss-120b4 |
| Context support | Groq lists full 131K context for GPT-OSS 120B.2 | SambaNova lists full 131K context for GPT-OSS 120B.4 |
When each option fits
Methodology and sources
For GPT-OSS 120B, AA performance uses 10,000 input tokens and high reasoning: output speed and first reasoning-chunk latency are medians. Endpoint Accuracy Index comes from a separate AA evaluation relative to a reference endpoint; it is not a pass rate. The uncached cost example bills 10,000 input and 1,000 output tokens at the named sellers’ standard tariffs. Rolling third-party performance values are the batch snapshot identified by the page’s Updated line, not live guarantees.
Build with Novita Model APIs
Create an API key and connect your application.