llmrun

DeepSeek V4 Flash 0731 vs Inkling Small

DeepSeek V4 Flash 0731 leads llmrun Score 127.6 to 119.6; neither fits 24 GB.

Model A · DeepSeek

DeepSeek V4 Flash 0731
127.6llmrun Score

304.2B · 27.4B active · 1M context · MIT · Jul 2026

Chat

Model B · Thinking Machines

Inkling Small
119.6llmrun Score

266.0B · 1.7B active · Apache 2.0 · Jul 2026

VisionChatFunctions

Which should you choose?

Choose DeepSeek V4 Flash 0731 if you want…

  • Higher llmrun Score

    127.6 vs 119.6

  • Stronger at math

    128.1 vs 117.1

  • Stronger at reasoning

    128.7 vs 122.6

  • Ahead on the benchmarks

    10 of 11 shared boards

Choose Inkling Small if you want…

  • Fits a Mac 128 GB

    IQ3_S · 116.2 GB; DeepSeek V4 Flash 0731 needs 84.1 GB at its smallest quant

  • Also does vision, function-calling

    DeepSeek V4 Flash 0731 doesn't list it

On your hardware

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

8 GB

RTX 4060 · 8 GB tier

DeepSeek V4 Flash 0731

Doesn't fit

smallest quant needs 84.1 GB

Inkling Small

Doesn't fit

smallest quant needs 76.3 GB

12 GB

RTX 3060 12GB · 12 GB tier

DeepSeek V4 Flash 0731

Doesn't fit

smallest quant needs 84.1 GB

Inkling Small

Doesn't fit

smallest quant needs 76.3 GB

16 GB

RTX 5060 Ti 16GB · 16 GB tier

DeepSeek V4 Flash 0731

Doesn't fit

smallest quant needs 84.1 GB

Inkling Small

Doesn't fit

smallest quant needs 76.3 GB

24 GBreference

RTX 4090 · 24 GB tier

DeepSeek V4 Flash 0731

Doesn't fit

smallest quant needs 84.1 GB

Inkling Small

Doesn't fit

smallest quant needs 76.3 GB

32 GB

RTX 5090 · 32 GB tier

DeepSeek V4 Flash 0731

Doesn't fit

smallest quant needs 84.1 GB

Inkling Small

Doesn't fit

smallest quant needs 76.3 GB

48 GB

RTX A6000 · 48 GB tier

DeepSeek V4 Flash 0731

Doesn't fit

smallest quant needs 84.1 GB

Inkling Small

Doesn't fit

smallest quant needs 76.3 GB

Mac 32 GB

Mac Mini M4 · Macs

DeepSeek V4 Flash 0731

Doesn't fit

smallest quant needs 84.1 GB

Inkling Small

Doesn't fit

smallest quant needs 76.3 GB

Mac 64 GB

Mac Studio M4 Max · Macs

DeepSeek V4 Flash 0731

Doesn't fit

smallest quant needs 84.1 GB

Inkling Small

Doesn't fit

smallest quant needs 76.3 GB

Mac 128 GB

Mac Studio M4 Max · Macs

DeepSeek V4 Flash 0731

Q2_K_S · ~27 tok/s

122.2 GB · up to 128K

Inkling Small

IQ3_S · ~97 tok/s

116.2 GB · up to 32Kbetter quant

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

Overall

DeepSeek V4 Flash 0731

127.6126–129

Inkling Small

119.6117–122

Coding

DeepSeek V4 Flash 0731

120.8114–127

Inkling Small

no data

Math

DeepSeek V4 Flash 0731

128.1126–131

Inkling Small

117.1113–121

Science

DeepSeek V4 Flash 0731

124.3112–136

Inkling Small

118.3108–129

Reasoning

DeepSeek V4 Flash 0731

128.7126–132

Inkling Small

122.6119–126

DeepSeek V4 Flash 0731 · #7 of 100 open models · #2 of 78 that fit ≤192 GB

Inkling Small · #13 of 100 open models · #5 of 78 that fit ≤192 GB

Benchmarks

11 shared boards: DeepSeek V4 Flash 0731 leads 10, Inkling Small leads 0 · Benchmarks imported Oct 2026

GPQA Diamond

DeepSeek V4 Flash 0731

91.0%#4/98 open

IndependentEpoch AI · Jul 2026

Inkling Small

88.5%#13/98 open

IndependentEpoch AI · Jul 2026

AIME 2024/2025

DeepSeek V4 Flash 0731

94.4%#8/86 open

IndependentEpoch AI · Jul 2026

Inkling Small

90.0%#14/86 open

IndependentEpoch AI · Jul 2026

FrontierMath

DeepSeek V4 Flash 0731

57.5%#7/17 open

IndependentEpoch AI · Jul 2026

Inkling Small

46.3%#11/17 open

IndependentEpoch AI · Jul 2026

ARC-AGI-2

DeepSeek V4 Flash 0731

61.4%#3/19 open

IndependentEpoch AI · Jul 2026

Inkling Small

40.1%#7/19 open

IndependentEpoch AI · Jul 2026

SciCode

DeepSeek V4 Flash 0731

49.9%#13/66 open

Lab-reportedEpoch AI · Jul 2026

Inkling Small

49.6%#14/66 open

Lab-reportedEpoch AI · Jul 2026

ARC-AGI

DeepSeek V4 Flash 0731

89.0%#5/19 open

IndependentEpoch AI · Jul 2026

Inkling Small

84.0%#7/19 open

IndependentEpoch AI · Jul 2026

SimpleQA

DeepSeek V4 Flash 0731

33.6%#14/17 open

IndependentEpoch AI · Jul 2026

Inkling Small

19.1%#16/17 open

IndependentEpoch AI · Jul 2026

FrontierMath Tier 4

DeepSeek V4 Flash 0731

24.4%#8/13 open

IndependentEpoch AI · Jul 2026

Inkling Small

17.1%#9/13 open

IndependentEpoch AI · Jul 2026

CritPt

DeepSeek V4 Flash 0731

16.6%#6/66 open

Lab-reportedEpoch AI · Jul 2026

Inkling Small

8.3%#12/66 open

Lab-reportedEpoch AI · Jul 2026

Mystery Game Puzzles

DeepSeek V4 Flash 0731

34.0%#3/18 open

IndependentEpoch AI · Jul 2026

Inkling Small

6.0%#17/18 open

IndependentEpoch AI · Jul 2026

13 more boards (other shared and single-model boards)
ProofBench

DeepSeek V4 Flash 0731

56.0%#5/26 open

Lab-reportedEpoch AI · Jul 2026

Inkling Small

6.0%#19/26 open

Lab-reportedEpoch AI · Jul 2026

Only DeepSeek V4 Flash 0731 (10)

Only Inkling Small (2)

VRAM by quantization

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

Q2_K

DeepSeek V4 Flash 0731

130 GB117.3 GB file

32K 130 GB · 128K 131 GB

Inkling Small

114 GB~113.0 GB file

32K 119 GB · 128K 136 GB

Q3_K_M

DeepSeek V4 Flash 0731

149 GB128.1 GB file

32K 149 GB · 128K 150 GB

Inkling Small

130 GB119.4 GB file

32K 136 GB · 128K 153 GB

Q4_K_M

DeepSeek V4 Flash 0731

183 GB~182.5 GB file

32K 183 GB · 128K 184 GB

Inkling Small

160 GB162.5 GB file

32K 166 GB · 128K 182 GB

Q5_K_M

DeepSeek V4 Flash 0731

217 GB~216.7 GB file

32K 217 GB · 128K 218 GB

Inkling Small

190 GB196.1 GB file

32K 195 GB · 128K 212 GB

Q6_K

DeepSeek V4 Flash 0731

251 GB~250.9 GB file

32K 252 GB · 128K 253 GB

Inkling Small

220 GB218.9 GB file

32K 225 GB · 128K 242 GB

Q8_0

DeepSeek V4 Flash 0731

305 GB~304.2 GB file

32K 305 GB · 128K 306 GB

Inkling Small

267 GB280.3 GB file

32K 272 GB · 128K 289 GB

14 more quants

IQ2_XXS

DeepSeek V4 Flash 0731

84.0 GB90.9 GB file

32K 84.3 GB · 128K 85.4 GB

Inkling Small

73.8 GB82.3 GB file

32K 79.1 GB · 128K 96.0 GB

IQ2_XS

DeepSeek V4 Flash 0731

91.6 GB88.1 GB file

32K 91.9 GB · 128K 93.0 GB

Inkling Small

not listed

IQ2_S

DeepSeek V4 Flash 0731

95.4 GB93.4 GB file

32K 95.7 GB · 128K 96.8 GB

Inkling Small

not listed

IQ2_M

DeepSeek V4 Flash 0731

103 GB90.9 GB file

32K 103 GB · 128K 104 GB

Inkling Small

90.4 GB82.4 GB file

32K 95.7 GB · 128K 113 GB

IQ3_XXS

DeepSeek V4 Flash 0731

118 GB104.2 GB file

32K 119 GB · 128K 120 GB

Inkling Small

104 GB97.9 GB file

32K 109 GB · 128K 126 GB

Q2_K_S

DeepSeek V4 Flash 0731

122 GB98.6 GB file

32K 122 GB · 128K 123 GB

Inkling Small

not listed

IQ3_XS

DeepSeek V4 Flash 0731

126 GB118.2 GB file

32K 126 GB · 128K 127 GB

Inkling Small

not listed

IQ3_S

DeepSeek V4 Flash 0731

130 GB116.1 GB file

32K 130 GB · 128K 131 GB

Inkling Small

114 GB107.5 GB file

32K 119 GB · 128K 136 GB

IQ3_M

DeepSeek V4 Flash 0731

137 GB135.8 GB file

32K 138 GB · 128K 139 GB

Inkling Small

not listed

IQ4_XS

DeepSeek V4 Flash 0731

164 GB136.7 GB file

32K 164 GB · 128K 165 GB

Inkling Small

144 GB127.4 GB file

32K 149 GB · 128K 166 GB

IQ4_NL

DeepSeek V4 Flash 0731

171 GB136.7 GB file

32K 172 GB · 128K 173 GB

Inkling Small

150 GB130.1 GB file

32K 156 GB · 128K 172 GB

Q4_K_S

DeepSeek V4 Flash 0731

not listed

Inkling Small

150 GB152.3 GB file

32K 156 GB · 128K 172 GB

Q5_K_S

DeepSeek V4 Flash 0731

not listed

Inkling Small

184 GB184.2 GB file

32K 189 GB · 128K 206 GB

BF16

DeepSeek V4 Flash 0731

609 GB~608.4 GB file

32K 609 GB · 128K 610 GB

Inkling Small

533 GB527.4 GB file

32K 538 GB · 128K 555 GB

How to run each

Specifications

Specifications of DeepSeek V4 Flash 0731 and Inkling Small
SpecDeepSeek V4 Flash 0731Inkling Small
PublisherDeepSeekThinking Machines
Parameters304.2B266.0B
Active parameters27B per token1.7B per token
ArchitectureDeepseekV4ForCausalLMInklingForConditionalGeneration
Context1M tokens—
ModalitiesChatVision, Chat, Function-calling
LicenseMITApache 2.0
ReleasedJul 2026Jul 2026

Frequently Asked Questions

Which is better, DeepSeek V4 Flash 0731 or Inkling Small?

DeepSeek V4 Flash 0731 leads llmrun Score 127.6 to 119.6; neither fits 24 GB. Choose DeepSeek V4 Flash 0731 for higher llmrun Score and stronger at math; choose Inkling Small for fits a Mac 128 GB and also does vision, function-calling.

Which needs less VRAM, DeepSeek V4 Flash 0731 or Inkling Small?

At Q4_K_M, DeepSeek V4 Flash 0731 requires 182.8 GB while Inkling Small needs 160.2 GB, making Inkling Small the easier model to run on your local setup.

Can I run DeepSeek V4 Flash 0731 or Inkling Small on a 16 GB GPU?

DeepSeek V4 Flash 0731 does not fit (smallest quant needs 84.1 GB); Inkling Small does not fit (smallest quant needs 76.3 GB).

DeepSeek V4 Flash 0731 vs Inkling Small — VRAM & GPU Comparison | llmrun