What is DeepSeek V4 Pro good for?

DeepSeek's flagship for the widest range of text tasks, with MIT weights.

When is DeepSeek V4 Pro not a good fit?

Cost-sensitive or self-hosted use. It is the most expensive in DeepSeek's lineup and needs about 960 GB to self-host at 4-bit. Text only.

What is the cheapest way to run DeepSeek V4 Pro?

Hosted, unless you push serious volume. DeepSeek charges $1.32 in / $3.96 out per 1M tokens. The cheapest rental that fits is 4x B300 at $22,723 a month, which costs the same as about 13B tokens a month on that API.

Can I self-host DeepSeek V4 Pro?

Yes. The weights are MIT licensed. At 4-bit it needs about 890 GB of GPU memory, which starts at roughly $22,723 a month on the cheapest rental that fits. See the table above for 8-bit and 16-bit.

Open weights · MIT

DeepSeek V4 Pro

DeepSeek's flagship 1.6T-parameter MoE (49B active) with hybrid sparse attention, previewed April 2026 and replaced by the 0813 release in August 2026, with weights under the MIT license.

Visit website Compare vs other LLMs →

Jump to: Specs Pricing Self-hosting Similar models FAQ More from DeepSeek

List price via DeepSeek

Input: $1.32 / 1M tokens
Output: $3.96 / 1M tokens

Cheapest: $0.45 / $0.89 per 1M tokens via GPUhub

6 providers

Key Specifications

Context window: 1M tokens
Max output: 384K tokens
Released: Apr 2026
Parameters: 1.6T, 49B active
Inputs: Text

Capabilities Show details: Reasoning Thinks before it answers, always on or as a switchable mode. Function calling Connect to external tools, APIs, and systems. Structured output Return responses in structured formats like JSON. Web search Search the web for up-to-date information.

Hosted API pricing

Input tokens M Output tokens M

Provider	Input / 1M tokens	Output / 1M tokens	Cached input / 1M	Cost at 10M in + 2M out
GPUhub Cheapest	$0.45	$0.89		$6.28	View
DeepSeek Creator	$1.32	$3.96		$21.12	View
Together AI	$1.32	$3.96	$0.13	$21.12	View
Geodd	$1.48	$3.40		$21.60	View
Novita	$1.60	$3.20		$22.40	View
Lyceum	$1.75	$3.50		$24.50	View

Heads up: Prices are estimates using base-tier, on-demand rates per 1M tokens from published pages; cached, batch and long-context tiers are shown only where listed. Providers may serve a shorter context or a quantized build than the creator's release. Verify with the provider before provisioning. How we estimate costs.

Estimated cost to self-host

DeepSeek V4 Pro needs about 890 GB of GPU memory at 4-bit. The cheapest rental that fits is 4× B300 at about $22,723 a month.

That costs the same as roughly 13B tokens a month on DeepSeek's API. Self-hosting is more expensive below that volume.

Precision	Cheapest, 1x concurrency	Cheapest, 8x concurrency
4-bit	4× B300 890 GB · $22,723/mo	4× B300 946 GB · $22,723/mo
8-bit	8× B300 1,610 GB · $45,446/mo	8× B300 1,666 GB · $45,446/mo
16-bit	Needs 3,210 GB	Needs 3,266 GB

Estimates, not quotes. Nvidia cards at median on-demand rates, 32K context per request. Break-even assumes 5:1 input to output. How we estimate costs.

Similarly priced models

The models nearest DeepSeek V4 Pro by blended rate, five each way, each at its own cheapest provider.

Model	Blended / 1M	Input / 1M	Output / 1M	Context	Cutoff	vs DeepSeek V4 Pro
Qwen3.6-35B-A3B Alibaba Cloud	$0.46	$0.25	$1.49	262K		−13%
Gemini 3.1 Flash-Lite Google Cloud	$0.46	$0.25	$1.50	1M	Jan 2025	−12%
Qwen3.6-Flash Alibaba Cloud	$0.46	$0.25	$1.50	1M		−12%
Qwen3.7-Plus Alibaba Cloud	$0.48	$0.32	$1.28	1M		−8%
Qwen3-VL-235B-A22B Instruct Alibaba Cloud	$0.50	$0.30	$1.50	262K		−4%
DeepSeek V4 Pro This model DeepSeek	$0.52	$0.45	$0.89	1M
GPT-5 Mini OpenAI	$0.54	$0.25	$2.00	400K	May 2024	+4%
Muse Glimmer 30B Meta	$0.54	$0.35	$1.50	131K	Jan 2026	+4%
Qwen3.5-35B-A3B Alibaba Cloud	$0.54	$0.25	$2.00	262K		+4%
DeepSeek V3 Turbo DeepSeek	$0.55	$0.40	$1.30	64K		+5%
Qwen3-Coder-480B-A35B Alibaba Cloud	$0.58	$0.38	$1.55	262K		+10%

Prices are USD per 1M tokens at each model's cheapest listed provider. Blended is the cost of 10M input plus 2M output tokens, spread over the 12M.

Packet·ai

Our sponsor

B200 @ $3.75/h on demand

No contract, no spot interrupts, no cluster minimums

Full root SSH, web dashboard, CI/CD ready

Reserve Now

Frequently Asked Questions

What is DeepSeek V4 Pro good for?: DeepSeek's flagship for the widest range of text tasks, with MIT weights.
When is DeepSeek V4 Pro not a good fit?: Cost-sensitive or self-hosted use. It is the most expensive in DeepSeek's lineup and needs about 960 GB to self-host at 4-bit. Text only.
What is the cheapest way to run DeepSeek V4 Pro?: Hosted, unless you push serious volume. DeepSeek charges $1.32 in / $3.96 out per 1M tokens. The cheapest rental that fits is 4x B300 at $22,723 a month, which costs the same as about 13B tokens a month on that API.
Can I self-host DeepSeek V4 Pro?: Yes. The weights are MIT licensed. At 4-bit it needs about 890 GB of GPU memory, which starts at roughly $22,723 a month on the cheapest rental that fits. See the table above for 8-bit and 16-bit.

More from DeepSeek

Model	Context	Input / 1M	Output / 1M
DeepSeek V3 Turbo Open weights	64K	$0.40	$1.30
DeepSeek V4 Flash Vision Exp	1M	$0.44	$1.32
DeepSeek V3 Open weights	128K	$0.27	$1.12
DeepSeek V3.1 Open weights	131K	$0.27	$1.00
DeepSeek V3.1 Terminus Open weights	131K	$0.27	$1.00
DeepSeek R1 Distill Llama 70B Open weights	131K	$0.80	$0.80
DeepSeek V3.2 Open weights	131K	$0.27	$0.40
DeepSeek R1 Turbo Open weights	64K	$0.70	$2.50