DeepSeek V4 Pro 0813 vs Kimi K3

Kimi K3 leads llmrun Score 133.6 to 129.2; neither fits 24 GB.

Model A · DeepSeek

DeepSeek V4 Pro 0813

129.2llmrun Score

1650.5B · 102.3B active · 1M context · MIT · Aug 2026

Chat

Model B · Moonshot AI

Kimi K3

133.6llmrun Score

2779.9B · 105.8B active · 1M context · Other · Jun 2026

VisionChat

Which should you choose?

Choose DeepSeek V4 Pro 0813 if you want…

Lighter on memory
990.6 vs 1673.7 GB at Q4_K_M
More permissive license
MIT vs Other

Choose Kimi K3 if you want…

Higher llmrun Score
133.6 vs 129.2
Stronger at coding
138.7 vs 125.2
Stronger at science
135.6 vs 126.3
Ahead on the benchmarks
14 of 21 shared boards
Also does vision
DeepSeek V4 Pro 0813 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 Pro 0813Kimi K3

8 GB

RTX 4060 · 8 GB tier

DeepSeek V4 Pro 0813

Doesn't fit

smallest quant needs 640.1 GB

Kimi K3

Doesn't fit

smallest quant needs 766.5 GB

12 GB

RTX 3060 12GB · 12 GB tier

DeepSeek V4 Pro 0813

Doesn't fit

smallest quant needs 640.1 GB

Kimi K3

Doesn't fit

smallest quant needs 766.5 GB

16 GB

RTX 5060 Ti 16GB · 16 GB tier

DeepSeek V4 Pro 0813

Doesn't fit

smallest quant needs 640.1 GB

Kimi K3

Doesn't fit

smallest quant needs 766.5 GB

24 GBreference

RTX 4090 · 24 GB tier

DeepSeek V4 Pro 0813

Doesn't fit

smallest quant needs 640.1 GB

Kimi K3

Doesn't fit

smallest quant needs 766.5 GB

32 GB

RTX 5090 · 32 GB tier

DeepSeek V4 Pro 0813

Doesn't fit

smallest quant needs 640.1 GB

Kimi K3

Doesn't fit

smallest quant needs 766.5 GB

48 GB

RTX A6000 · 48 GB tier

DeepSeek V4 Pro 0813

Doesn't fit

smallest quant needs 640.1 GB

Kimi K3

Doesn't fit

smallest quant needs 766.5 GB

Mac 32 GB

Mac Mini M4 · Macs

DeepSeek V4 Pro 0813

Doesn't fit

smallest quant needs 640.1 GB

Kimi K3

Doesn't fit

smallest quant needs 766.5 GB

Mac 64 GB

Mac Studio M4 Max · Macs

DeepSeek V4 Pro 0813

Doesn't fit

smallest quant needs 640.1 GB

Kimi K3

Doesn't fit

smallest quant needs 766.5 GB

Mac 128 GB

Mac Studio M4 Max · Macs

DeepSeek V4 Pro 0813

Doesn't fit

smallest quant needs 640.1 GB

Kimi K3

Doesn't fit

smallest quant needs 766.5 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 0813 Kimi K3

Overall

DeepSeek V4 Pro 0813

129.2127–131

Kimi K3

133.6132–135

Coding

DeepSeek V4 Pro 0813

125.2119–132

Kimi K3

138.7133–145

Agents

DeepSeek V4 Pro 0813

no data

Kimi K3

132.9128–138

Math

DeepSeek V4 Pro 0813

128.3126–131

Kimi K3

134.7132–137

Science

DeepSeek V4 Pro 0813

126.3114–138

Kimi K3

135.6124–147

Reasoning

DeepSeek V4 Pro 0813

131.2128–134

Kimi K3

129.5127–133

DeepSeek V4 Pro 0813 · #5 of 107 open models · #5 of 106 that fit ≤1 TB

Kimi K3 · #2 of 107 open models · #2 of 106 that fit ≤2 TB

Benchmarks

21 shared boards: DeepSeek V4 Pro 0813 leads 7, Kimi K3 leads 14 · Benchmarks imported Oct 2026

BoardDeepSeek V4 Pro 0813Kimi K3

GPQA Diamond

DeepSeek V4 Pro 0813

91.7%#3/98 open

IndependentEpoch AI · Aug 2026

Kimi K3

93.1%#1/98 open

IndependentEpoch AI · Jul 2026

LiveBench Coding

DeepSeek V4 Pro 0813

77.2#7/19 open

IndependentLiveBench · Jun 2026

Kimi K3

81.5#1/19 open

IndependentLiveBench · Jun 2026

LiveBench Reasoning

DeepSeek V4 Pro 0813

85.8#6/19 open

IndependentLiveBench · Jun 2026

Kimi K3

90.7#1/19 open

IndependentLiveBench · Jun 2026

AIME 2024/2025

DeepSeek V4 Pro 0813

98.6%#1/86 open

IndependentEpoch AI · Aug 2026

Kimi K3

97.2%#4/86 open

IndependentEpoch AI · Jul 2026

FrontierMath

DeepSeek V4 Pro 0813

64.6%#5/17 open

IndependentEpoch AI · Aug 2026

Kimi K3

72.2%#1/17 open

IndependentEpoch AI · Jul 2026

LiveBench Math

DeepSeek V4 Pro 0813

95.1#1/19 open

IndependentLiveBench · Jun 2026

Kimi K3

84.4#13/19 open

IndependentLiveBench · Jun 2026

ARC-AGI-2

DeepSeek V4 Pro 0813

61.3%#4/19 open

IndependentEpoch AI · Aug 2026

Kimi K3

60.4%#5/19 open

IndependentEpoch AI · Jul 2026

SciCode

DeepSeek V4 Pro 0813

51.0%#9/66 open

Lab-reportedEpoch AI · Aug 2026

Kimi K3

59.5%#2/66 open

Lab-reportedEpoch AI · Jul 2026

ARC-AGI

DeepSeek V4 Pro 0813

90.5%#4/19 open

IndependentEpoch AI · Aug 2026

Kimi K3

94.5%#2/19 open

IndependentEpoch AI · Jul 2026

SimpleQA

DeepSeek V4 Pro 0813

52.9%#2/17 open

IndependentEpoch AI · Aug 2026

Kimi K3

50.6%#3/17 open

IndependentEpoch AI · Jul 2026

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

FrontierMath Tier 4

DeepSeek V4 Pro 0813

26.8%#6/13 open

IndependentEpoch AI · Aug 2026

Kimi K3

39.0%#1/13 open

IndependentEpoch AI · Jul 2026

WeirdML

DeepSeek V4 Pro 0813

66.2%#4/46 open

IndependentEpoch AI · Aug 2026

Kimi K3

82.6%#1/46 open

IndependentEpoch AI · Jul 2026

FrontierCode

DeepSeek V4 Pro 0813

28.5%#6/12 open

Lab-reportedEpoch AI · Aug 2026

Kimi K3

44.2%#1/12 open

Lab-reportedEpoch AI · Jul 2026

ALE-Bench

DeepSeek V4 Pro 0813

1403.2#2/43 open

Lab-reportedEpoch AI · Aug 2026

Kimi K3

1524.5#1/43 open

Lab-reportedEpoch AI · Jul 2026

APEX-Agents

DeepSeek V4 Pro 0813

47.3%#6/16 open

Lab-reportedEpoch AI · Aug 2026

Kimi K3

50.6%#4/16 open

Lab-reportedEpoch AI · Jul 2026

Chess Puzzles

DeepSeek V4 Pro 0813

47.0%#1/67 open

IndependentEpoch AI · Aug 2026

Kimi K3

39.0%#2/67 open

IndependentEpoch AI · Jul 2026

CritPt

DeepSeek V4 Pro 0813

18.0%#5/66 open

Lab-reportedEpoch AI · Aug 2026

Kimi K3

23.4%#2/66 open

Lab-reportedEpoch AI · Jul 2026

DTBench

DeepSeek V4 Pro 0813

93.9%#1/68 open

Lab-reportedEpoch AI · Aug 2026

Kimi K3

91.2%#3/68 open

Lab-reportedEpoch AI · Jul 2026

LMCA

DeepSeek V4 Pro 0813

45.5%#5/54 open

Lab-reportedEpoch AI · Aug 2026

Kimi K3

52.7%#2/54 open

Lab-reportedEpoch AI · Jul 2026

Mystery Game Puzzles

DeepSeek V4 Pro 0813

43.0%#2/18 open

IndependentEpoch AI · Aug 2026

Kimi K3

26.0%#5/18 open

IndependentEpoch AI · Jul 2026

ProofBench

DeepSeek V4 Pro 0813

50.0%#7/26 open

Lab-reportedEpoch AI · Aug 2026

Kimi K3

87.0%#1/26 open

Lab-reportedEpoch AI · Jul 2026

LMArena Text

DeepSeek V4 Pro 0813

1464.6#9/187 open

ArenaLMArena · Oct 2026

Kimi K3

1487.1#1/187 open

ArenaLMArena · Oct 2026

LMArena WebDev

DeepSeek V4 Pro 0813

1582.0#11/52 open

ArenaLMArena · Oct 2026

Kimi K3

1655.2#1/52 open

ArenaLMArena · Oct 2026

Only Kimi K3 (7)

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 Pro 0813Kimi K3

Q2_K

DeepSeek V4 Pro 0813

702 GB587.3 GB file

32K 702 GB · 128K 704 GB

Kimi K3

1187 GB928.6 GB file

32K 1185 GB · 128K 1196 GB

Q3_K_M

DeepSeek V4 Pro 0813

805 GB763.7 GB file

32K 805 GB · 128K 807 GB

Kimi K3

1361 GB~1355.2 GB file

32K 1359 GB · 128K 1370 GB

Q4_0

DeepSeek V4 Pro 0813

826 GB885.3 GB file

32K 826 GB · 128K 827 GB

Kimi K3

not listed

Q4_K_M

DeepSeek V4 Pro 0813

991 GB~990.3 GB file

32K 991 GB · 128K 992 GB

Kimi K3

1674 GB~1668.0 GB file

32K 1672 GB · 128K 1682 GB

Q5_K_M

DeepSeek V4 Pro 0813

1176 GB~1176.0 GB file

32K 1177 GB · 128K 1178 GB

Kimi K3

1986 GB~1980.7 GB file

32K 1985 GB · 128K 1995 GB

Q6_K

DeepSeek V4 Pro 0813

1362 GB~1361.7 GB file

32K 1362 GB · 128K 1364 GB

Kimi K3

2299 GB~2293.4 GB file

32K 2297 GB · 128K 2308 GB

Q8_0

DeepSeek V4 Pro 0813

1651 GB~1650.5 GB file

32K 1651 GB · 128K 1653 GB

Kimi K3

2786 GB2955.0 GB file

32K 2784 GB · 128K 2794 GB

4 more quants

IQ2_XXS

DeepSeek V4 Pro 0813

not listed

Kimi K3

770 GB711.1 GB file

32K 768 GB · 128K 779 GB

IQ3_XXS

DeepSeek V4 Pro 0813

640 GB619.5 GB file

32K 640 GB · 128K 642 GB

Kimi K3

not listed

Q4_1

DeepSeek V4 Pro 0813

929 GB983.6 GB file

32K 929 GB · 128K 930 GB

Kimi K3

not listed

BF16

DeepSeek V4 Pro 0813

3301 GB~3301.0 GB file

32K 3302 GB · 128K 3303 GB

Kimi K3

5566 GB~5559.9 GB file

32K 5564 GB · 128K 5574 GB

How to run each

DeepSeek V4 Pro 0813

Ollama · Q4_K_M

$ollama run hf.co/unsloth/DeepSeek-V4-Pro-0813-GGUF

All ways to run DeepSeek V4 Pro 0813 →

Kimi K3

Ollama · Q4_K_M

$ollama run hf.co/unsloth/Kimi-K3-GGUF

All ways to run Kimi K3 →

Specifications

Specifications of DeepSeek V4 Pro 0813 and Kimi K3
Spec	DeepSeek V4 Pro 0813	Kimi K3
Publisher	DeepSeek	Moonshot AI
Parameters	1650.5B	2779.9B
Active parameters	102B per token	106B per token
Architecture	DeepseekV4ForCausalLM	KimiK3ForConditionalGeneration
Context	1M tokens	1M tokens
Modalities	Chat	Vision, Chat
License	MIT	Other
Released	Aug 2026	Jun 2026

Frequently Asked Questions

Which is better, DeepSeek V4 Pro 0813 or Kimi K3?

Kimi K3 leads llmrun Score 133.6 to 129.2; neither fits 24 GB. Choose DeepSeek V4 Pro 0813 for lighter on memory and more permissive license; choose Kimi K3 for higher llmrun Score and stronger at coding.

Which needs less VRAM, DeepSeek V4 Pro 0813 or Kimi K3?

At Q4_K_M, DeepSeek V4 Pro 0813 requires 990.6 GB while Kimi K3 needs 1673.7 GB, making DeepSeek V4 Pro 0813 the easier model to run on your local setup.

Can I run DeepSeek V4 Pro 0813 or Kimi K3 on a 16 GB GPU?

DeepSeek V4 Pro 0813 does not fit (smallest quant needs 640.1 GB); Kimi K3 does not fit (smallest quant needs 766.5 GB).

Which has a longer context window, DeepSeek V4 Pro 0813 or Kimi K3?

DeepSeek V4 Pro 0813 supports 1,048,576 tokens and Kimi K3 supports 1,048,576 tokens.