← All models
NVIDIA: Nemotron 3 Super
nvidia/nemotron-3-super-120b-a12btextReasoning
NVIDIA open 120B model activating just 12B parameters (hybrid Mamba-Transformer MoE) — high compute efficiency and accuracy in complex multi-agent applications. Adjustable reasoning depth, 1M context.
Pricing
Request / 1M
Response / 1M
Parameters
messagesrequiredreasoning_effortoff · low · mediumtemperature0–2max_tokensoptionalstreamtrue · falseHow to use via API
from openai import OpenAIclient = OpenAI(base_url="https://api.mixen.ai/v1", api_key="mxn-...")resp = client.chat.completions.create(model="nvidia/nemotron-3-super-120b-a12b",messages=[{"role": "user", "content": "Привет"}],reasoning_effort="low", # off · low · mediumtemperature=0.7,max_tokens=1024,stream=False,)print(resp.choices[0].message.content)# Стриминг: stream=True — ответ приходит чанкамиfor chunk in client.chat.completions.create(model="nvidia/nemotron-3-super-120b-a12b",messages=[{"role": "user", "content": "Привет"}],stream=True,):print(chunk.choices[0].delta.content or "", end="", flush=True)
