NVIDIA: Nemotron 3 Ultra

nvidia/nemotron-3-ultra-550b-a55b

NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it...

Prompt $/1M

$0.600

Completion $/1M

$3.600

Context

1,000,000 tok

Best provider

181 tok/s · BaseTen

Specs

Model id: nvidia/nemotron-3-ultra-550b-a55b
Display name: NVIDIA: Nemotron 3 Ultra
Modality: text->text
Tokenizer: Other
Instruct type: —
Context length: 1,000,000
Max completion: —
Prompt $/1M: $0.600
Completion $/1M: $3.600
Free tier: no

Capabilities

Function callingno
JSON modeno
System promptno
Visionno
Moderatedno

Benchmark scores

Each axis is one benchmark, normalized to the best model in the field.

No benchmark data for this model.

Speed by provider

Provider	tok/s	TTFT	Total latency
BaseTenfastest	181	493 ms	754 ms
Nebiusfastest	137	776 ms	862 ms
Together	130	944 ms	1,582 ms
Crusoe	85	594 ms	1,128 ms
DeepInfra	69	3,471 ms	4,731 ms
Venice	39	1,783 ms	3,957 ms