AI model comparison

Nemotron 3 Ultra vs DeepSeek V4 Pro 0423

Side by side comparison of Nemotron 3 Ultra and DeepSeek V4 Pro 0423 on pricing, benchmark scores, context window, and more.

NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE).

Input / 1M
$0.50
Output / 1M
$2.20
Context
262K

DeepSeek V4 Pro is a large-scale Mixture-of-Experts model from DeepSeek with 1.6T total parameters and 49B activated parameters, supporting a 1M-token context window.

Input / 1M
$0.90
Output / 1M
$1.50
Context
1.05M

Comparing image generators? Compare image models

At a glance

Quick verdict

How the two models stack up on the things people ask about most.

Lower price

nvidia logo

Nemotron 3 Ultra

1.2x cheaper for a typical chat

Higher intelligence score

deepseek logo

DeepSeek V4 Pro 0423

20.8 vs 11.8 on Artificial Analysis

Larger context window

deepseek logo

DeepSeek V4 Pro 0423

1.05M vs 262K tokens

Newer release

nvidia logo

Nemotron 3 Ultra

Released June 4, 2026

Benchmarks

Benchmark scores

Independent scores from Artificial Analysis. Higher is better.

Intelligence index

nvidia logoNemotron 3 Ultra11.8
deepseek logoDeepSeek V4 Pro 042320.8

Coding index

nvidia logoNemotron 3 Ultra49.3
deepseek logoDeepSeek V4 Pro 042359.4

Agentic index

nvidia logoNemotron 3 Ultra20.1
deepseek logoDeepSeek V4 Pro 042326.3

Pricing

Nemotron 3 Ultra vs DeepSeek V4 Pro 0423 API pricing

Per-token API rates. Cheaper option highlighted.

Metricnvidia logoNemotron 3 Ultradeepseek logoDeepSeek V4 Pro 0423
Input tokensPer 1M tokens$0.50$0.90
Output tokensPer 1M tokens$2.20$1.50
Cached input (read)Per 1M tokens$0.10$0.20
Cache writePer 1M tokensNot availableNot available

What it costs in practice

Estimated cost per 1,000 requests at standard rates. Coding and long-document figures also show the cost when the input is already cached.

Chat message

2K in, 500 out

nvidia logoNemotron 3 Ultra$2.10
deepseek logoDeepSeek V4 Pro 0423$2.55

Coding task

30K in, 4K out

nvidia logoNemotron 3 Ultra$23.80$11.80 cached
deepseek logoDeepSeek V4 Pro 0423$33$12 cached

Long document summary

150K in, 2K out

nvidia logoNemotron 3 Ultra$79.40$19.40 cached
deepseek logoDeepSeek V4 Pro 0423$138$33 cached

Specs

Context window and capabilities

How much each model can read, how much it can write, and what it accepts as input.

Metricnvidia logoNemotron 3 Ultradeepseek logoDeepSeek V4 Pro 0423
Context window262K tokens1.05M tokens
Max output16K tokens393K tokens
Input typesTextText
Output typesTextText
Reasoning effort levelsHigh, Medium, NoneExtra high, High, None
Default reasoning effortHighHigh
Release dateJune 4, 2026April 24, 2026

Features

Supported features

API features available for each model.

Metricnvidia logoNemotron 3 Ultradeepseek logoDeepSeek V4 Pro 0423
Tool callingSupportedSupported
Structured outputsSupportedSupported
JSON modeSupportedSupported
ReasoningSupportedSupported
TemperatureSupportedSupported
Stop sequencesSupportedSupported
Deterministic seedSupportedSupported
Verbosity controlNot supportedNot supported

Our take

Nemotron 3 Ultra is 1.2x cheaper for a typical chat. DeepSeek V4 Pro 0423 scores higher on the Artificial Analysis Intelligence Index (20.8 vs 11.8). DeepSeek V4 Pro 0423 has the larger context window (1.05M vs 262K tokens). The right pick depends on your workload, so the quickest way to settle it is to send the same prompt to both and compare.

FAQ

Frequently asked questions

Is Nemotron 3 Ultra or DeepSeek V4 Pro 0423 cheaper?

Nemotron 3 Ultra is cheaper for a typical chat (2K input tokens and 500 output tokens). Nemotron 3 Ultra costs $0.50 per 1M input tokens and $2.20 per 1M output tokens, while DeepSeek V4 Pro 0423 costs $0.90 per 1M input tokens and $1.50 per 1M output tokens.

Which has a bigger context window, Nemotron 3 Ultra or DeepSeek V4 Pro 0423?

DeepSeek V4 Pro 0423 supports up to 1.05M tokens of context, compared with 262K for Nemotron 3 Ultra.

Is Nemotron 3 Ultra smarter than DeepSeek V4 Pro 0423?

On the Artificial Analysis Intelligence Index, Nemotron 3 Ultra scores 11.8 and DeepSeek V4 Pro 0423 scores 20.8, putting DeepSeek V4 Pro 0423 ahead. Benchmarks don't capture everything, so the best test is running your own prompts through both.

Can I use Nemotron 3 Ultra and DeepSeek V4 Pro 0423 at the same time?

Yes. Shortcut Chat sends one prompt to multiple models at once, so you can see Nemotron 3 Ultra and DeepSeek V4 Pro 0423 answer side by side and keep the better response.

nvidia logodeepseek logo

Why choose? Ask both.

Send one prompt to Nemotron 3 Ultra and DeepSeek V4 Pro 0423 at the same time. Compare the answers side by side and keep the best one.