DeepSeek V4 Pro.
Open-weights 1.6T Mixture-of-Experts text model with 49B active parameters and a one million token context, the larger DeepSeek-V4 preview, under the MIT licence.
4× AMD MI325 · $12.38/hr.
Most-aggressive quantisation we have a working recommendation for. Lower precision = less VRAM = cheaper hardware, at a small accuracy cost.
Cheapest hosted endpoints.
Variants in the DeepSeek family.
DeepSeek's flagship MoE — 671B total, 37B active, frontier-class.
DeepSeek's reasoning model — RL-trained, frontier-class, MIT-licensed.
70B Llama distilled from DeepSeek R1's reasoning traces.
32B Qwen base distilled from DeepSeek R1.
14B distilled R1 — laptop-friendly reasoning.
7B distilled R1 — runs on any modern GPU.
Tiny distilled R1 — phone / browser deployable.
Official release of DeepSeek-V4-Pro that supersedes the preview, adding a spe...
DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, wi...
DeepSeek-V3.1 is post-trained on the top of DeepSeek-V3.1-Base, which is buil...
DeepSeek V3, a 685B-parameter, mixture-of-experts model, is the latest iterat...
DeepSeek-V3.1 Terminus is an update to [DeepSeek V3.1](/deepseek/deepseek-cha...
DeepSeek-V3.2 is a large language model designed to harmonize high computatio...
DeepSeek-V3.2-Speciale is a high-compute variant of DeepSeek-V3.2 optimized f...
Open-weights 284B Mixture-of-Experts text model with 13B active parameters an...
May 28th update to the [original DeepSeek R1](/deepseek/deepseek-r1) Performa...
Open-weights multimodal 552B Mixture-of-Experts model that reads images and t...
Minor update to DeepSeek-V4.1-Flash with improved reasoning, coding, agent an...
DeepSeek-V4-Flash-Vision-Exp is an experimental MIT-licensed model adding ima...
DeepSeek-V3.2-Exp is an experimental large language model released by DeepSee...
Frequently asked.
How do I run DeepSeek V4 Pro?
Where can I access DeepSeek V4 Pro?
How much does it cost to run DeepSeek V4 Pro?
Is DeepSeek V4 Pro open-source or proprietary?
Cheapest hardware per quantisation.
Each row is one quantisation tier (the same weights compressed differently). Lower precision → lower VRAM → cheaper hardware, at the cost of small accuracy loss. $/hr is the cheapest rate on offer now. A multi-card setup uses a real instance of that size when one is rented; otherwise it is the single-card rate times the count.
| Quantisation | Cheapest GPU config | Total VRAM | Live $/hr | tokens/sec | |
|---|---|---|---|---|---|
|
FP8
FP8 — 8-bit float (Hopper / Blackwell)
|
2048 GB |
$24.51/hr
Cyfuture AI · on-demand
|
— | Compare → | |
|
INT4
INT4 — 4-bit integer (~4× VRAM saving)
|
1024 GB |
$12.38/hr
Cyfuture AI · on-demand
|
— | Compare → |
What it costs per month across providers.
Estimate your monthly bill for DeepSeek V4 Pro across every host that publishes per-token pricing. Slide your token volumes; the chart + table re-rank cheapest-first.
Cheapest provider on the left.
Total monthly cost — input + output tokens combined.
Bill breakdown.
Rent the GPU instead of paying per token.
For an open-weights model like DeepSeek V4 Pro, you can rent a GPU and serve inference yourself. The math: cheapest GPU rental × 730 hours/month. Power is included in a rental's hourly price.
Assumes the GPU runs 24/7 at ~85% utilisation. If your traffic is bursty, you'll pay less for the API and probably more for the GPU (idle hours still cost rental). The breakeven analysis lives on the Self-host vs API breakeven tool.
About DeepSeek V4 Pro.
The larger model in the DeepSeek-V4 preview pair, a Mixture-of-Experts language model with 1.6T total parameters and 49B active per token. It supports a one million token context and uses a hybrid attention design that DeepSeek says needs about a quarter of the per-token compute and a tenth of the cache of its V3.2 predecessor at that length. It offers a maximum reasoning effort mode and is released under the MIT licence.