NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 vs DeepSeek-V4-Pro-0813

Specs, VRAM requirements and download trends — updated 24 October 2026.

nvidia · A · Open weights

Params

560.5 B

Context

262,144

30d

1.1 M

deepseek-ai · B · Open weights

Params

1,650.5 B

Context

1,048,576

30d

150 K

mit · commercial OK

Specification comparison

Differences are highlighted; identical values are muted.

Specification	NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16	DeepSeek-V4-Pro-0813
Deployment	Open weights · also via API	Open weights · also via API
Parameters	560.5 B	1,650.5 B
Architecture	NemotronHForCausalLM	DeepseekV4ForCausalLM
Context length	262,144	1,048,576
Licence	other	mit
Commercial use	Unknown	Allowed
Languages	12	—
Downloads 30d	1,067,855	150,373
Downloads all time	2.1 M	380 K
Quantizations on the Hub	SAFETENSORS	SAFETENSORS
Gated	No	No
First seen on the Hub	2026-06-03	2026-08-13
API input price / 1M tokens	$0.60	$0.66
API output price / 1M tokens	$2.40	$1.98
Est. cost per 1M input + 1M output tokens	$3.00	$2.64
Cached input / 1M tokens	$0.12	$0.022
AA Intelligence Index	22.9	36
AA Coding Index	49.3	68.8
AA Agentic Index	20.1	41.3
Max output tokens	182,520	393,216
API providers	3	21
Tool calling	Yes	Yes
Reasoning mode	Yes	Yes