SambaNova
Llama and DeepSeek at high speed on custom hardware.
SambaNova Cloud runs seven models on its own RDU chips: Llama 3.3 70B, DeepSeek V3.1 and V3.2, MiniMax M2.7 and M3, Gemma 4 31B and GPT-OSS 120B. Fast when it answers; the free ceiling is tight — 20 requests a minute, 20 a day, 200K tokens a day.
Email, no card. MiniMax M3 takes the window to 1M. R1 left the catalogue in April 2026. OpenAI-compatible. Use it when you want those weights on custom silicon, not when you need volume.
- Free models
- 7
- Max context
- 1M
- Free tier
- 20 requests per day
- Requirement
- Email account, no card
Endpoint
Base URL
https://api.sambanova.ai/v1Env var
SAMBANOVA_API_KEYRate limit
20 requests/minute, 20 requests/day, 200K tokens/day- OpenAI-compatible
- text
- image
- reasoning
Code examples
npm install openaiimport OpenAI from 'openai'
const client = new OpenAI({
baseURL: 'https://api.sambanova.ai/v1',
apiKey: process.env.SAMBANOVA_API_KEY
})
const response = await client.chat.completions.create({
model: 'DeepSeek-V3.1',
messages: [{ role: 'user', content: 'Explain closures in one paragraph.' }]
})
console.log(response.choices[0].message.content)pip install openaiimport os
from openai import OpenAI
client = OpenAI(
base_url="https://api.sambanova.ai/v1",
api_key=os.environ["SAMBANOVA_API_KEY"],
)
response = client.chat.completions.create(
model="DeepSeek-V3.1",
messages=[{"role": "user", "content": "Explain closures in one paragraph."}],
)
print(response.choices[0].message.content)curl https://api.sambanova.ai/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $SAMBANOVA_API_KEY" \
-d '{
"model": "DeepSeek-V3.1",
"messages": [{"role": "user", "content": "Explain closures in one paragraph."}]
}'Models you get for free7
- Llama 3.3 70B
Meta-Llama-3.3-70B-Instruct - DeepSeek V3.1
DeepSeek-V3.1 - DeepSeek V3.2
DeepSeek-V3.2 - MiniMax M2.7
MiniMax-M2.7 - MiniMax M3
MiniMax-M3 - Gemma 4 31B
Gemma-4-31B - GPT-OSS 120B
gpt-oss-120b
