Nemotron 3 Ultra

wandb/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B · family nemotron · price provider-declared (via models.dev), USD per 1M tokens. Capabilities: toolsreasoningtext.

$1.25Blended $/1M (3:1)

$0.75Input $/1M

$2.75Output $/1M

$3.501M in + 1M out

$0.15 −80%Cache read $/1M

#3428 of 5,000Cheapest rank

Provider: Weights & Biases (wandb)
Context: 262,144 tokens
Max output: 262,144 tokens
Capabilities: toolsreasoningtext

Same model, other providers

“Nemotron 3 Ultra” is offered by 3 providers — cheapest-effective first. Vercel AI Gateway is the cheapest at $1.05 blended.

Provider	Input	Output	Blended	Context
Vercel AI Gateway	$0.6	$2.4	$1.05	1000k
DigitalOcean	$0.9	$1.7	$1.10	131k
Weights & Biases (this)	$0.75	$2.75	$1.25	262k

Call it

jishie indexes & prices models; it does not proxy inference. Most providers are OpenAI-compatible — point base_url at Weights & Biases and pass this model id:

curl $BASE_URL/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" -H "Content-Type: application/json" \
  -d '{"model":"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B","messages":[{"role":"user","content":"hello"}]}'

Route programmatically with MCP find_model / get_model, or fetch this record free (no wallet):