OpenModelsHQ

DeepSeek-V4-Flash-0731

DeepSeek-V4-Flash-0731 is an open-weight language model from DeepSeek with 304B parameters (mixture of experts). At 4-bit it needs about 201.6 GB of memory, so the smallest device in our directory that fits it is the Rubin, which is announced but not shipping yet.

Overview

Estimated memory at 4-bit
201.6GB

Includes room for an 8k context and runtime overhead. More context needs more memory.

Total parameters
304B
Mixture of experts, all weights counted
Context window
1024K
1,048,576 tokens
Cheapest API, blended / 1M
$0.0875
₹8.25 · Relace · 3:1 input to output
All DeepSeek API prices

Key information

Lab
DeepSeek
Released
31 Jul 2026
Parameters
304B · MoE
Type
Language
Context
1,048,576 tokens
License
MIT
HF downloads, 30 days
2,366,650
View all

Will it run on your hardware?

How we estimate
Memory needed vs devices
DeepSeek-V4-Flash-0731 needs about 201.6 GB at 4-bit, 355.5 GB at 8-bit and 670 GB at 16-bit. At 4-bit it fits none of the reference devices. Dashed lines are usable memory on each device; a model fits when its estimate stays within 90% of it. Log scale. 8k context.
Memory needed vs devices: data
Estimate or deviceMemory
Q4201.6 GB
Q8355.5 GB
BF16670 GB
RTX 5060 Ti 16 GB16 GB usable
RTX 5090 32 GB32 GB usable
DGX Spark 128 GB108.8 GB usable
Mac Studio M5 Ultra 256 GB192 GB usable

Single-device memory estimate. Pick a device to update the On your device column.

36 GB usable memory

FormatMemory neededOn your deviceExample hardware
4-bit (Q4)≈ 201.6 GBToo big 560%RubinAnnounced288 GB memory
8-bit (Q8)≈ 355.5 GBToo big 988%No single device in our directory has enough memory
16-bit (BF16)≈ 670 GBToo big 1861%No single device in our directory has enough memory

Fits means the estimate stays within 90% of usable memory. Tight is above 90% and up to 100%. Too big is above 100%. These are memory checks, not speed or compatibility tests.

Mac allowances are 66% of total RAM up to 36 GB, and 75% above. AI boxes allow 85% because their memory is shared with the operating system.

Where to run it

On a Mac

No trusted MLX build found yet.

Check your Mac

With GGUF

Choose a file for llama.cpp or a compatible app.

unsloth/DeepSeek-V4-Flash-0731-GGUFFind compatible hardware

Through an API

Relace (fp4) is the cheapest provider we found.

$0.07 / $0.14

Input / output per 1M tokens
₹6.60 / ₹13 at ₹94.29 per $1

Blended: $0.0875 (₹8.25) per 1M. Ranked by a 3:1 mix of input and output tokens.

GGUF with llama.cppView repository
llama-cli -hf unsloth/DeepSeek-V4-Flash-0731-GGUF \
  -c 8192 -p "Explain how a transformer works."

Example commands built from the repository ids above. They need a working install of llama.cpp. We have not run them; check each repository for available files.

Hosted API 24

USD and INR per million tokensAll DeepSeek API prices
Precision
RelaceCheapest$0.0875₹8.25$0.07₹6.60$0.14₹131,048,576FP4
Baidu$0.0897₹8.46$0.0598₹5.64$0.1795₹171,048,576FP8
DeepInfra$0.09₹8.49$0.06₹5.66$0.18₹171,048,576FP8
Wafer$0.1076₹10$0.0434₹4.09$0.30₹281,048,576NVFP4
DigitalOcean$0.1487₹14$0.119₹11$0.238₹221,048,576Not listed
BaseTen$0.1625₹15$0.13₹12$0.26₹251,048,576FP32
BaseTen$0.1625₹15$0.13₹12$0.26₹251,048,576FP8
Venice$0.1625₹15$0.13₹12$0.26₹251,000,000Not listed
CoreWeave$0.165₹16$0.13₹12$0.27₹25262,144FP8
Cohere$0.175₹17$0.14₹13$0.28₹261,048,576FP8
Parasail$0.175₹17$0.14₹13$0.28₹261,048,576FP8
Reka$0.18₹17$0.10₹9.43$0.42₹40262,144Not listed
Sail Research$0.1875₹18$0.15₹14$0.30₹281,048,576FP4
StreamLake$0.198₹19$0.132₹12$0.396₹371,024,000FP8
Inceptron$0.2213₹21$0.13₹12$0.495₹471,048,576FP4
Mancer 2$0.2375₹22$0.15₹14$0.50₹471,048,576FP8
SiliconFlow$0.33₹31$0.22₹21$0.66₹621,048,576FP8
GMICloud$0.429₹40$0.286₹27$0.858₹811,048,575FP8
Alibaba$0.528₹50$0.352₹33$1.056₹1001,000,000Not listed
Novita$0.6138₹58$0.4092₹39$1.2276₹1161,048,576FP8
AtlasCloud$0.66₹62$0.44₹41$1.32₹1241,048,576FP4
Chutes$0.66₹62$0.44₹41$1.32₹1241,048,576FP8
Cloudflare$0.66₹62$0.44₹41$1.32₹1241,048,576Not listed
PrimeIntellect$0.66₹62$0.44₹41$1.32₹1241,048,576FP4
View on OpenRouterEndpoint sourceSorted by the blended price: 3 parts input to 1 part output. Context limits differ by provider.

INR at ₹94.29 per $1. Prices fetched 14 Nov 2026.

More about this model

Frequently asked questions

How much memory does DeepSeek-V4-Flash-0731 need? At an 8k context, about 201.6 GB at 4-bit, 355.5 GB at 8-bit and 670 GB at 16-bit (BF16). A longer context needs more.

Can I run DeepSeek-V4-Flash-0731 on a Mac? No current Mac fits the 4-bit estimate of 201.6 GB. The smallest Mac in our directory that does is the Mac Studio M3 Ultra (512GB), which is discontinued. We have not found a trusted MLX build yet, but a GGUF build is available.

Does the full context fit? DeepSeek-V4-Flash-0731 supports up to 1,048,576 tokens of context. Our memory estimates use 8k, so long prompts need more memory than shown.

What does it cost through an API? The cheapest provider we found is Relace, at $0.07 per 1M input tokens and $0.14 per 1M output tokens (₹6.60 and ₹13 at 94.29 rupees per dollar). Ranked by a 3:1 mix of input and output tokens.

License in plain English

MIT. This license permits commercial use, modification and redistribution, as long as the copyright and license notice are kept. This summary is not legal advice.

Read MIT

Read about this license

OMHQ RigBench

Not yet tested. We show no speed or quality score without a measurement.

Revision history and sources

Last updated . Sources:Hugging Face model card, GGUF repository, OpenRouter.