Models / DeepSeek V4 Flash 0731

DeepSeek V4 Flash 0731

MoE workstation

284B total, 13B active per token (MoE). Same FP4+FP8 hybrid-attention family as V4 Pro: Compressed Sparse Attention (CSA) + Heavily Compressed Attention (HCA) across 61 layers, manifold-constrained Hyper-Connections (mHC), Muon optimizer, 32T+ pretraining tokens.

  • Context: 1M native, 384K max output; three modes (non-think / think-high / think-max).

The 2026-07-31 iterative update of V4 Flash (supersedes the April model). Per Artificial Analysis, a 10-point Intelligence Index jump to 50 - 6 points above V4 Pro, 1 behind GLM 5.2 / GPT-5.6 Luna, 7 behind Kimi K3. Agentic Elo 1559 (up from 1189), Terminal-Bench 2.1 79% (+17), Humanity’s Last Exam 37% (+5), GPQA-Diamond 91% (+1), SciCode 50% (+5). Token usage -12%; hallucination rate 84% (a 12-point drop); AA-Omniscience Index -16 (+7). Pricing unchanged at $0.14/$0.28 per 1M in/out (cache-hit $0.0028/M, a 98% discount). The model the deepseek-chat/reasoner API aliases now route to (retired 2026-07-24).

  • Local run (2026-07-31): Unsloth’s Dynamic 2.0 GGUFs landed local inference. UD-Q4_K_XL is a 155GB lossless 4-bit build (~168GB RAM); UD-Q8_K_XL is a 162GB 8-bit full-precision build (~175GB RAM, only 7GB bigger than Q4 because the 13B active experts dominate). Both fit two stacked DGX Sparks (256GB unified via ConnectX-7) or a 192GB+ unified rig; a smaller 3-bit (~110GB RAM) that would fit a single 128GB Spark is announced but not yet published. Run via Unsloth or llama.cpp -hf - Ollama only ships the :cloud endpoint, so there is no local Ollama tag.

Open weights under MIT (full weights expected in the coming weeks per DeepSeek; the Unsloth GGUFs are available now).

Agents on Rails benchmark (Aug 2026, Le Mans round). 65.1% accuracy on 63 runs - last of the 16-model field, with only 12.7% API recall. The low recall suggests it often hand-rolled fixes instead of using the relevant Rails API. Still the cheapest per-token model in the field, so the gap is cost-vs-accuracy rather than capability at any price.

AI-generated content marks

This model embeds text watermarks in generated text and adds C2PA provenance metadata to supported files such as .png, .jpg, and .svg. Marks can be lost through editing, screenshots, or format conversion, so their absence does not prove a file is human-made.

Provider transparency docs

coding reasoning agentic
Parameters
284.0B
Context
1000k
License
mit
Developer
DeepSeek
Origin
🇨🇳 China
Released
Jul 2026

What people are building with DeepSeek V4 Flash 0731

Real demos from X

Ben Davis's 2x NVIDIA DGX Spark rack hits 50+ tok/s on DeepSeek V4 Flash 0731 with multiple sessions View on X →

Benchmark scores

Vendor-reported - from the developer's own model card / tech report

Agents' Last Exam
25.2
Automation-Bench
37.7
CyberGym
76.7
DeepSWE
54.4
ExploitGym
1.8
GPQA-Diamond
89.9
HLE
37.8
HLE (w/ tools)
51.5
MathArena Apex
58.6
NL2Repo-Bench
54.2
SEC-Bench Pro
30.9
Terminal-Bench 2.1
82.7
Terminal-Bench 3.0
7.6
Terminal-Bench 4.0
7.0

Vendor-reported - from the developer's own model card / tech report

Ran this model on your own hardware? Join free and add your measured tok/s to the community numbers.

Score per dollar

3000 pts per $/M input

general_score (90) divided by cheapest input price ($0.03/M). Higher is better value. See live pricing.

Related models

Guides covering DeepSeek V4 Flash 0731

Save your hardware and every model page answers the real question: will it run on your machine, and how fast?

Join free - save your rig →

Run it locally

Per-quant memory needs and a static "can you run it?" reference - no rig entry required

UD-Q4_K_XL
155.0GB 168.0GB min 192.0GB rec
Unsloth Dynamic 4-bit XL - lossless at Q4 size
UD-Q8_K_XL
162.0GB 175.0GB min 192.0GB rec
Unsloth Dynamic 8-bit XL - full precision lossless

The reference hardware

Schematic of the NVIDIA Jetson Orin NX 16GB reference rig - 16GB unified memory, 102 GB/s aggregate bandwidth
Schematic of the Jetson AGX Orin 64GB reference rig - 64GB unified memory, 204 GB/s aggregate bandwidth
Schematic of the Single GTX 1080 Ti (11GB) reference rig - 11GB VRAM, 484 GB/s aggregate bandwidth
Schematic of the Single RTX 4090 (24GB) reference rig - 24GB VRAM, 1008 GB/s aggregate bandwidth
Schematic of the Single RTX 5090 (32GB) reference rig - 32GB VRAM, 1792 GB/s aggregate bandwidth
Schematic of the RTX PRO 6000 Blackwell (96GB) reference rig - 96GB VRAM, 1792 GB/s aggregate bandwidth
Schematic of the 2x RTX 3090 (48GB) reference rig - 48GB VRAM, 1872 GB/s aggregate bandwidth
Schematic of the 2x RTX 5090 (64GB) reference rig - 64GB VRAM, 3584 GB/s aggregate bandwidth
Schematic of the 4x RTX 4090 (96GB) reference rig - 96GB VRAM, 4032 GB/s aggregate bandwidth
Schematic of the 4x RTX 5090 (128GB) reference rig - 128GB VRAM, 7168 GB/s aggregate bandwidth
Schematic of the MacBook Pro M5 Max 128GB reference rig - 128GB unified memory, 614 GB/s aggregate bandwidth
Schematic of the Ryzen AI Max+ 395 128GB reference rig - 128GB unified memory, 256 GB/s aggregate bandwidth
Schematic of the DGX Spark 128GB unified reference rig - 128GB unified memory, 273 GB/s aggregate bandwidth
Schematic of the 4x H100 80GB (320GB) reference rig - 320GB VRAM, 13400 GB/s aggregate bandwidth
Schematic of the NVIDIA DGX Station 748GB reference rig - 748GB unified memory, 8000 GB/s aggregate bandwidth
Schematic of the 8x RTX 3090 rack (192GB) reference rig - 192GB VRAM, 7489 GB/s aggregate bandwidth
Schematic of the AMD Instinct MI300X (192GB) reference rig - 192GB VRAM, 5324 GB/s aggregate bandwidth
Schematic of the Mac Studio M4 Ultra 192GB reference rig - 192GB unified memory, 1092 GB/s aggregate bandwidth
Schematic of the Mac Studio M4 Ultra 512GB reference rig - 512GB unified memory, 1092 GB/s aggregate bandwidth
Schematic of the Dual EPYC 9004 + 768GB DDR5-4800 reference rig - 768GB unified memory, 460 GB/s aggregate bandwidth
Schematic of the Epyc + 512GB DDR4-3200 + 2x RTX 3090 reference rig - 560GB unified memory, 204 GB/s aggregate bandwidth
Schematic of the Epyc + 512GB DDR4-2400 + 2x RTX 3090 reference rig - 560GB unified memory, 153 GB/s aggregate bandwidth

22 reference configs, drawn in-house. Scroll for more.

Can you run it? - reference rigs

Rig UD-Q4_K_XL UD-Q8_K_XL
NVIDIA Jetson Orin NX 16GB no -> cloud no -> cloud
Jetson AGX Orin 64GB no -> cloud no -> cloud
Single GTX 1080 Ti (11GB) no -> cloud no -> cloud
Single RTX 4090 (24GB) no -> cloud no -> cloud
Single RTX 5090 (32GB) no -> cloud no -> cloud
RTX PRO 6000 Blackwell (96GB) no -> cloud no -> cloud
2x RTX 3090 (48GB) no -> cloud no -> cloud
2x RTX 5090 (64GB) no -> cloud no -> cloud
4x RTX 4090 (96GB) no -> cloud no -> cloud
4x RTX 5090 (128GB) offload offload
MacBook Pro M5 Max 128GB no -> cloud no -> cloud
Ryzen AI Max+ 395 128GB no -> cloud no -> cloud
DGX Spark 128GB unified no -> cloud no -> cloud
4x H100 80GB (320GB) fast 2454.2t/s fast 993.1t/s
NVIDIA DGX Station 748GB fast 1465.2t/s fast 592.9t/s
8x RTX 3090 rack (192GB) fast 1371.7t/s fast 555.1t/s
AMD Instinct MI300X (192GB) fast 412.4t/s fast 394.6t/s
Mac Studio M4 Ultra 192GB fast 92.3t/s fast 88.3t/s
Mac Studio M4 Ultra 512GB fast 92.3t/s fast 88.3t/s
Dual EPYC 9004 + 768GB DDR5-4800 fast 35.7t/s fast 34.2t/s
Epyc + 512GB DDR4-3200 + 2x RTX 3090 ok 15.9t/s ok 15.2t/s
Epyc + 512GB DDR4-2400 + 2x RTX 3090 ok 11.9t/s ok 11.4t/s

Fit tiers use the same will-it-run logic as the rig finder. For comfortable fits, the badge reflects decode speed: fast >=20 t/s, ok 8-20 t/s, slow <8 t/s. t/s is a bandwidth estimate, not a measured benchmark.

Download options

UD-Q4_K_XL Unsloth - local-optimized -1% vs fp16
155.0GB dl 168.0GB min 192.0GB rec
REC RAM vs largest quant
155GB lossless 4-bit weights (Unsloth UD-Q4_K_XL) + ~13GB overhead + KV cache; ~168GB to run in RAM
UD-Q8_K_XL Unsloth - local-optimized -1% vs fp16
162.0GB dl 175.0GB min 192.0GB rec
REC RAM vs largest quant
162GB 8-bit full-precision weights (Unsloth UD-Q8_K_XL) + ~13GB overhead + KV cache; ~175GB to run in RAM

Or run it in the cloud

Live per-provider pricing, throughput and uptime - refreshed about 19 hours ago via OpenRouter. Click a column to sort.

some pricing may be stale - last verified 2026-09-17

Provider Type Input $/M Output $/M Cache $/M Tok/s Latency Uptime Value
OpenInference
API 0.03 0.13 0.010 - - 100.00% best uptime
DigitalOcean stale
API 0.08 0.25 0.025 - - -
Nous Portal stale
API 0.11 0.22 - - - -
BaseTen
API 0.13 0.26 0.028 - - 100.00%
CoreWeave
API 0.13 0.28 0.070 - - 100.00%
Io Net stale
API 0.22 0.49 0.114 - - 100.00%
Alibaba
API 0.35 1.06 0.035 - - 100.00%
Novita
API 0.41 1.23 0.026 - - 100.00%
Cloudflare
API 0.44 1.32 0.014 - - 100.00%
DeepSeek
API 0.44 1.32 0.014 - - 100.00%
AtlasCloud
API 0.44 1.32 0.028 - - 100.00%
GMICloud
API 0.29 0.86 0.009 - - 99.99%
Baidu
API 0.44 1.32 0.014 - - 99.98%
Wafer
API 0.10 0.25 0.050 - - 99.95%
Makora
API 0.09 0.20 0.020 - - 99.93%
Together
API 0.14 0.28 0.030 - - 99.93%
DeepInfra
API 0.06 0.18 0.015 - - 99.92%
Reka
API 0.11 0.66 0.007 - - 99.92%
DigitalOcean
API 0.12 0.24 0.024 - - 99.92%
Venice
API 0.18 0.35 0.035 - - 99.87%
Relace
API 0.06 0.12 0.012 - - 99.82%
SiliconFlow
API 0.22 0.66 0.028 - - 99.77%
NextBit
API 0.35 1.06 0.012 - - 99.77%
Morph
API 0.14 0.40 0.036 - - 99.69%
Inceptron
API 0.06 0.20 0.010 - - 99.68%
Phala
API 0.44 1.32 0.028 - - 99.54%
Fireworks
API 0.22 0.66 0.007 - - 99.36%
Sail Research
API 0.07 0.34 0.023 - - 99.22%
StreamLake
API 0.06 0.18 0.002 - - 99.19%
Ionstream stale
API 0.21 0.42 0.100 - - 98.56%
Parasail
API 0.14 0.28 0.050 - - 98.17%
AkashML
API 0.06 0.18 0.016 - - 97.35%
Ambient risky
API 0.08 0.18 0.016 - - 94.57%
Nebius risky stale
API 0.14 0.28 0.016 - - 91.05%
Decart avoid stale
API 0.06 0.13 0.013 - - 89.81%
Mancer 2 avoid
API 0.20 0.60 0.012 - - 84.82%

Default order: throughput among 95%+ uptime providers, then latency; subscriptions last. Sort by any column. Subscription rows show $/mo in the Value column - per-token columns are "-". Affiliate links are marked sponsored / nofollow. Confirm current pricing on the provider's site before committing.

Detailed API pricing page + JSON endpoint →

See who runs DeepSeek in production →

PRICE HISTORY

Inference cost over time

Data accumulates from the first daily sync - longer ranges populate over time. Prices come from OpenRouter snapshots, not a historical API.

Loading price history...