DeepSeek V4 Pro vs Inkling

DeepSeek V4 Pro and Inkling match in llmrun Score, with values of 116.5 versus 116.0; neither fits 24 GB.

Model A · DeepSeek

DeepSeek V4 Pro

116.5llmrun Score

1598.8B · 50.7B active · 1M context · MIT · Apr 2026

Chat

Model B · Thinking Machines

Inkling

116.0llmrun Score

952.4B · 18.1B active · Apache 2.0 · Jul 2026

VisionChatFunctions

Which should you choose?

Choose DeepSeek V4 Pro if you want…

Ahead on the benchmarks
14 of 17 shared boards

Choose Inkling if you want…

Lighter on memory
572.1 vs 959.6 GB at Q4_K_M
Also does vision, function-calling
DeepSeek V4 Pro doesn't list it

On your hardware

Each model at its default quant (Q4_K_M), a smaller one when it must. Estimates, ±20%.

MemoryDeepSeek V4 ProInkling

8 GB

RTX 4060 · 8 GB tier

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Inkling

Doesn't fit

smallest quant needs 408.4 GB

12 GB

RTX 3060 12GB · 12 GB tier

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Inkling

Doesn't fit

smallest quant needs 408.4 GB

16 GB

RTX 5060 Ti 16GB · 16 GB tier

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Inkling

Doesn't fit

smallest quant needs 408.4 GB

24 GBreference

RTX 4090 · 24 GB tier

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Inkling

Doesn't fit

smallest quant needs 408.4 GB

32 GB

RTX 5090 · 32 GB tier

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Inkling

Doesn't fit

smallest quant needs 408.4 GB

48 GB

RTX A6000 · 48 GB tier

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Inkling

Doesn't fit

smallest quant needs 408.4 GB

Mac 32 GB

Mac Mini M4 · Macs

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Inkling

Doesn't fit

smallest quant needs 408.4 GB

Mac 64 GB

Mac Studio M4 Max · Macs

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Inkling

Doesn't fit

smallest quant needs 408.4 GB

Mac 128 GB

Mac Studio M4 Max · Macs

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Inkling

Doesn't fit

smallest quant needs 408.4 GB

Reference GPU: RTX 4090. Tinted cell = the better pick on that machine (fits, more headroom, or at least 10% faster).

llmrun Score

with its 90% range · bars share one scale

DeepSeek V4 Pro Inkling

Overall

DeepSeek V4 Pro

116.5114–119

Inkling

116.0114–118

Coding

DeepSeek V4 Pro

110.6105–116

Inkling

110.2105–115

Agents

DeepSeek V4 Pro

no data

Inkling

111.4102–120

Math

DeepSeek V4 Pro

117.6114–122

Inkling

112.1107–117

Science

DeepSeek V4 Pro

122.3111–134

Inkling

116.4106–127

Reasoning

DeepSeek V4 Pro

117.0110–124

Inkling

121.0118–124

DeepSeek V4 Pro · #16 of 100 open models · #16 of 99 that fit ≤1 TB

Inkling · #18 of 100 open models · #15 of 96 that fit ≤640 GB

Benchmarks

17 shared boards: DeepSeek V4 Pro leads 14, Inkling leads 3 · Benchmarks imported Oct 2026

BoardDeepSeek V4 ProInkling

GPQA Diamond

DeepSeek V4 Pro

90.9%#5/98 open

IndependentEpoch AI · Apr 2026

Inkling

88.3%#14/98 open

IndependentEpoch AI · Jul 2026

LiveBench Coding

DeepSeek V4 Pro

70.0#16/19 open

IndependentLiveBench · Jun 2026

Inkling

71.0#14/19 open

IndependentLiveBench · Jun 2026

LiveBench Reasoning

DeepSeek V4 Pro

82.7#10/19 open

IndependentLiveBench · Jun 2026

Inkling

78.3#14/19 open

IndependentLiveBench · Jun 2026

AIME 2024/2025

DeepSeek V4 Pro

96.7%#5/86 open

IndependentEpoch AI · Apr 2026

Inkling

88.9%#17/86 open

IndependentEpoch AI · Jul 2026

FrontierMath

DeepSeek V4 Pro

45.3%#12/17 open

IndependentEpoch AI · Apr 2026

Inkling

33.3%#15/17 open

IndependentEpoch AI · Jul 2026

LiveBench Math

DeepSeek V4 Pro

90.7#3/19 open

IndependentLiveBench · Jun 2026

Inkling

88.4#7/19 open

IndependentLiveBench · Jun 2026

SciCode

DeepSeek V4 Pro

50.0%#12/66 open

Lab-reportedEpoch AI · Apr 2026

Inkling

47.0%#18/66 open

Lab-reportedEpoch AI · Jul 2026

SimpleQA

DeepSeek V4 Pro

47.0%#4/17 open

IndependentEpoch AI · Apr 2026

Inkling

40.3%#8/17 open

IndependentEpoch AI · Jul 2026

FrontierMath Tier 4

DeepSeek V4 Pro

2.4%#13/13 open

IndependentEpoch AI · Apr 2026

Inkling

4.9%#12/13 open

IndependentEpoch AI · Jul 2026

WeirdML

DeepSeek V4 Pro

48.9%#10/46 open

IndependentEpoch AI · Apr 2026

Inkling

32.3%#37/46 open

IndependentEpoch AI · Jul 2026

26 more boards (other shared and single-model boards)

FrontierCode

DeepSeek V4 Pro

17.6%#9/12 open

Lab-reportedEpoch AI · Apr 2026

Inkling

14.0%#11/12 open

Lab-reportedEpoch AI · Jul 2026

ALE-Bench

DeepSeek V4 Pro

1006.1#10/43 open

Lab-reportedEpoch AI · Apr 2026

Inkling

946.0#11/43 open

Lab-reportedEpoch AI · Jul 2026

Chess Puzzles

DeepSeek V4 Pro

19.0%#12/14 open

IndependentEpoch AI · Apr 2026

Inkling

21.0%#7/14 open

IndependentEpoch AI · Jul 2026

CritPt

DeepSeek V4 Pro

12.9%#9/66 open

Lab-reportedEpoch AI · Apr 2026

Inkling

5.4%#16/66 open

Lab-reportedEpoch AI · Jul 2026

DTBench

DeepSeek V4 Pro

90.7%#6/68 open

Lab-reportedEpoch AI · Apr 2026

Inkling

87.5%#13/68 open

Lab-reportedEpoch AI · Jul 2026

LMCA

DeepSeek V4 Pro

41.2%#8/54 open

Lab-reportedEpoch AI · Apr 2026

Inkling

37.6%#11/54 open

Lab-reportedEpoch AI · Jul 2026

ProofBench

DeepSeek V4 Pro

16.0%#14/26 open

Lab-reportedEpoch AI · Apr 2026

Inkling

0.0%#26/26 open

Lab-reportedEpoch AI · Jul 2026

LMArena Text

DeepSeek V4 Pro

1457.8#11/187 open

ArenaLMArena · Oct 2026

Inkling

1440.9#20/187 open

ArenaLMArena · Oct 2026

LMArena WebDev

DeepSeek V4 Pro

1464.2#20/52 open

ArenaLMArena · Oct 2026

Inkling

1413.7#26/52 open

ArenaLMArena · Oct 2026

Only DeepSeek V4 Pro (5)

Only Inkling (12)

VRAM by quantization

Memory to load the model and 2K context; below it, the total at 32K and 128K context. Lighter side in bold.

QuantDeepSeek V4 ProInkling

Q2_K

DeepSeek V4 Pro

680 GB574.1 GB file

32K 680 GB · 128K 682 GB

Inkling

405 GB~404.8 GB file

32K 412 GB · 128K 432 GB

Q3_K_M

DeepSeek V4 Pro

780 GB748.7 GB file

32K 780 GB · 128K 782 GB

Inkling

465 GB~464.3 GB file

32K 471 GB · 128K 491 GB

Q4_K_M

DeepSeek V4 Pro

960 GB954.6 GB file

32K 960 GB · 128K 961 GB

Inkling

572 GB~571.4 GB file

32K 578 GB · 128K 598 GB

Q5_K_M

DeepSeek V4 Pro

1140 GB~1139.2 GB file

32K 1140 GB · 128K 1141 GB

Inkling

679 GB~678.6 GB file

32K 686 GB · 128K 705 GB

Q6_K

DeepSeek V4 Pro

1319 GB~1319.0 GB file

32K 1320 GB · 128K 1321 GB

Inkling

786 GB~785.7 GB file

32K 793 GB · 128K 813 GB

Q8_0

DeepSeek V4 Pro

1599 GB1671.8 GB file

32K 1600 GB · 128K 1601 GB

Inkling

953 GB856.8 GB file

32K 959 GB · 128K 979 GB

1 more quant

BF16

DeepSeek V4 Pro

3198 GB~3197.7 GB file

32K 3198 GB · 128K 3200 GB

Inkling

1905 GB1894.3 GB file

32K 1912 GB · 128K 1932 GB

How to run each

DeepSeek V4 Pro

Ollama · Q4_K_M

$ollama run hf.co/teamblobfish/DeepSeek-V4-Pro-GGUF:Q4_K_M

All ways to run DeepSeek V4 Pro →

Inkling

Ollama · Q4_K_M

$ollama run hf.co/unsloth/inkling-GGUF

All ways to run Inkling →

Specifications

Specifications of DeepSeek V4 Pro and Inkling
Spec	DeepSeek V4 Pro	Inkling
Publisher	DeepSeek	Thinking Machines
Parameters	1598.8B	952.4B
Active parameters	51B per token	18B per token
Architecture	DeepseekV4ForCausalLM	InklingForConditionalGeneration
Context	1M tokens	—
Modalities	Chat	Vision, Chat, Function-calling
License	MIT	Apache 2.0
Released	Apr 2026	Jul 2026

Frequently Asked Questions

Which is better, DeepSeek V4 Pro or Inkling?

DeepSeek V4 Pro and Inkling match in llmrun Score, with values of 116.5 versus 116.0; neither fits 24 GB. Choose DeepSeek V4 Pro for ahead on the benchmarks; choose Inkling for lighter on memory and also does vision, function-calling.

Which needs less VRAM, DeepSeek V4 Pro or Inkling?

At Q4_K_M, DeepSeek V4 Pro requires 959.6 GB while Inkling needs 572.1 GB, making Inkling the easier model to run on your local setup.

Can I run DeepSeek V4 Pro or Inkling on a 16 GB GPU?

DeepSeek V4 Pro does not fit (smallest quant needs 680 GB); Inkling does not fit (smallest quant needs 408.4 GB).