DeepSeek

DeepSeek V4 Pro

Name: DeepSeek V4 Pro
Author: DeepSeek

DeepSeek V4 Pro is a DeepSeek model with 1M context, $1.32 in / $3.96 out per million tokens, last verified October 1, 2026. Open weights: yes. Status: Legacy since September 10, 2026; the current pick is DeepSeek V4.1 Flash. Public-board scores are linked to the publisher; vendor-only figures stay in the watch-out, not on the board.

DeepSeek V4 Pro is DeepSeek's large open-weight (MIT) MoE with a 1M-token context — DeepSeek says the cheaper V4.1-Flash now beats it.

Recorded profile verification October 1, 2026Individual provider fields may change

AI model specification details
Specification	DeepSeek V4 Pro
Provider	DeepSeek
Tier	Frontier
Context window	1M
Max output	384K
Input / 1M tokens	$1.32
Output / 1M tokens	$3.96
Weights	Open
Parameters	1.6T total / 49B active (MoE)
Reasoning levels	low, high, max
Modalities	text
License	MIT
API model id	deepseek-v4-pro
Released	April 24, 2026

Pricing tiers: Official peak rates $1.32/$3.96 per MTok; off-peak $0.66/$1.98 (cache-miss). Cache hit $0.044 peak / $0.022 off-peak. Peak hours are weekdays 01:00–04:00 and 06:00–10:00 UTC, excluding Chinese public holidays; weekends, holidays and all other hours are off-peak. Open weights (MIT). 1.6T/49B active MoE.

Publisher evidence

Published benchmark results

Each result keeps its source date visible. Dates refer to source publication or capture unless the source explicitly states when it measured the result. A missing benchmark is not treated as a zero.

Published benchmark results for DeepSeek V4 Pro
Benchmark	Score	Source date	Source
Artificial Analysis Intelligence Index [max]DeepSeek V4 Pro 0813 (max)	36	2026-10-01	Artificial Analysis · View source
SWE-bench Verified [max]Original April 24 checkpoint; historical vendor evaluation, not the served 0813 refresh	80.6%	2026-04-24	DeepSeek original V4 Pro HF card (vendor, Think Max; verifier/harness comparability unverified) · View source
GPQA Diamond [max]Original April 24 checkpoint; historical vendor evaluation, not the served 0813 refresh	90.1%	2026-04-24	DeepSeek V4 Pro HF model card (vendor, Think Max, pass@1) · View source
Humanity's Last Exam [max]Original April 24 checkpoint; historical vendor evaluation, not the served 0813 refresh	48.2%	2026-04-24	DeepSeek V4 Pro HF model card (vendor, Think Max, with tools; 37.7 no tools) · View source
MMLU-Pro [max]Original April 24 checkpoint; historical vendor evaluation, not the served 0813 refresh	87.5%	2026-04-24	DeepSeek V4 Pro HF model card (vendor, Think Max, exact match) · View source
Humanity's Last Exam0813 GA refresh; without tools; vendor report, exact run configuration unverified	42.7%	2026-08-13	DeepSeek August 13 GA changelog (vendor) · View source
Humanity's Last Exam0813 GA refresh; with tools; vendor report, exact run configuration unverified	60%	2026-08-13	DeepSeek August 13 GA changelog (vendor) · View source
Terminal-Bench 2.10813 GA refresh; vendor report, exact run configuration unverified	87.9%	2026-08-13	DeepSeek August 13 GA changelog (vendor) · View source
DeepSWE 1.10813 GA refresh; vendor report, exact run configuration unverified	62.7%	2026-08-13	DeepSeek August 13 GA changelog (vendor) · View source

Best for

Cost-sensitive hosted agents
Open-weight deployments
High-volume coding

Watch out

The current tariff identifies DeepSeek-V4-Pro-0813. DeepSeek's current changelog reverses the earlier September 14 re-route plan and says it will continue serving V4 Pro at unchanged rates until further notice. Original April benchmark/checkpoint results are not the current 0813 evaluation. Self-hosting memory depends on quantization, offloading and hardware.

History

DeepSeek V4 Pro timeline

Release and status events for this model that passed two-source verification — dated, sourced, and linked.

DeepSeek V4-Pro phase-out announcedStatus updateSeptember 10, 2026
DeepSeek says it is phasing out V4-Pro in favour of V4.1-Flash.
DeepSeek API changelog
DeepSeek-V4-ProGeneral availabilityAugust 13, 2026
GA checkpoint rolled out on app, web, and API. The API id remains `deepseek-v4-pro`. Peak/off-peak pricing starts 2026-08-16 16:00 UTC.
DeepSeek API changelog

Source receipts

Catalog figures for this model were checked against the following sources.

Whole-profile primary source check October 1 (accessed 2026-10-01)
Whole-profile primary source check October 1 (accessed 2026-10-01)
DeepSeek current API version, holiday tariff and 0813 GA benchmark publication (accessed 2026-10-01)
DeepSeek current tariff and served endpoint versions (accessed 2026-10-01)
DeepSeek — V4 news (accessed 2026-08-29)
DeepSeek API pricing (accessed 2026-08-29)
DeepSeek — V4.1-Flash release (V4-Pro phase-out) (accessed 2026-09-26)
DeepSeek API changelog (accessed 2026-09-26)

Common questions

DeepSeek V4 Pro

Answered from the verified figures on this page rather than general guidance.

How much does DeepSeek V4 Pro cost per million tokens?

DeepSeek V4 Pro lists base rates of $1.32 per million input tokens and $3.96 per million output tokens. Estimate both input and output token volume and any additional charges for your workload. Official peak rates $1.32/$3.96 per MTok; off-peak $0.66/$1.98 (cache-miss). Cache hit $0.044 peak / $0.022 off-peak. Peak hours are weekdays 01:00–04:00 and 06:00–10:00 UTC, excluding Chinese public holidays; weekends, holidays and all other hours are off-peak. Open weights (MIT). 1.6T/49B active MoE.

What is DeepSeek V4 Pro's context window?

DeepSeek V4 Pro accepts about 1M tokens of context. That only matters if you routinely send very long documents, large codebases, or multi-turn histories that approach that limit.

What is DeepSeek V4 Pro best for?

DeepSeek V4 Pro is a frontier tier from DeepSeek. It suits cost-sensitive hosted agents, open-weight deployments, high-volume coding. The current tariff identifies DeepSeek-V4-Pro-0813. DeepSeek's current changelog reverses the earlier September 14 re-route plan and says it will continue serving V4 Pro at unchanged rates until further notice. Original April benchmark/checkpoint results are not the current 0813 evaluation. Self-hosting memory depends on quantization, offloading and hardware.

Can I self-host DeepSeek V4 Pro?

DeepSeek V4 Pro publishes open weights, but self-hosting depends on the licence, hardware footprint, quantisation quality, and serving stack. A hosted API is often cheaper until you have measured throughput and concurrency on your own hardware.

Model comparison catalog Benchmark library