DeepSeek V4 Pro vs Kimi K2.7 Code

DeepSeek V4 Pro and Kimi K2.7 Code match in llmrun Score, with values of 116.5 versus 118.1; 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 · Moonshot AI

Kimi K2.7 Code

118.1llmrun Score

1026.9B · 33.3B active · 256K context · Other · Jun 2026

VisionCodeFunctions

Which should you choose?

Choose DeepSeek V4 Pro if you want…

Longer context window
1M vs 256K tokens
Also does chat
Kimi K2.7 Code doesn't list it
More permissive license
MIT vs Other

Choose Kimi K2.7 Code if you want…

Lighter on memory
620.0 vs 959.6 GB at Q4_K_M
Also does vision, code, 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 ProKimi K2.7 Code

8 GB

RTX 4060 · 8 GB tier

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Kimi K2.7 Code

Doesn't fit

smallest quant needs 283.8 GB

12 GB

RTX 3060 12GB · 12 GB tier

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Kimi K2.7 Code

Doesn't fit

smallest quant needs 283.8 GB

16 GB

RTX 5060 Ti 16GB · 16 GB tier

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Kimi K2.7 Code

Doesn't fit

smallest quant needs 283.8 GB

24 GBreference

RTX 4090 · 24 GB tier

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Kimi K2.7 Code

Doesn't fit

smallest quant needs 283.8 GB

32 GB

RTX 5090 · 32 GB tier

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Kimi K2.7 Code

Doesn't fit

smallest quant needs 283.8 GB

48 GB

RTX A6000 · 48 GB tier

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Kimi K2.7 Code

Doesn't fit

smallest quant needs 283.8 GB

Mac 32 GB

Mac Mini M4 · Macs

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Kimi K2.7 Code

Doesn't fit

smallest quant needs 283.8 GB

Mac 64 GB

Mac Studio M4 Max · Macs

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Kimi K2.7 Code

Doesn't fit

smallest quant needs 283.8 GB

Mac 128 GB

Mac Studio M4 Max · Macs

DeepSeek V4 Pro

Doesn't fit

smallest quant needs 680 GB

Kimi K2.7 Code

Doesn't fit

smallest quant needs 283.8 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 Kimi K2.7 Code

Overall

DeepSeek V4 Pro

116.5114–119

Kimi K2.7 Code

118.1115–121

Coding

DeepSeek V4 Pro

110.6105–116

Kimi K2.7 Code

114.8110–120

Agents

DeepSeek V4 Pro

no data

Kimi K2.7 Code

118.7112–125

Math

DeepSeek V4 Pro

117.6114–122

Kimi K2.7 Code

120.7116–125

Science

DeepSeek V4 Pro

122.3111–134

Kimi K2.7 Code

117.7107–128

Reasoning

DeepSeek V4 Pro

117.0110–124

Kimi K2.7 Code

115.4106–125

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

Kimi K2.7 Code · #14 of 100 open models · #12 of 96 that fit ≤640 GB

Benchmarks

16 shared boards: DeepSeek V4 Pro leads 8, Kimi K2.7 Code leads 7 · Benchmarks imported Oct 2026

BoardDeepSeek V4 ProKimi K2.7 Code

GPQA Diamond

DeepSeek V4 Pro

90.9%#5/98 open

IndependentEpoch AI · Apr 2026

Kimi K2.7 Code

87.9%#15/98 open

IndependentEpoch AI · Jun 2026

LiveBench Coding

DeepSeek V4 Pro

70.0#16/19 open

IndependentLiveBench · Jun 2026

Kimi K2.7 Code

74.0#11/19 open

IndependentLiveBench · Jun 2026

LiveBench Reasoning

DeepSeek V4 Pro

82.7#10/19 open

IndependentLiveBench · Jun 2026

Kimi K2.7 Code

82.8#9/19 open

IndependentLiveBench · Jun 2026

AIME 2024/2025

DeepSeek V4 Pro

96.7%#5/86 open

IndependentEpoch AI · Apr 2026

Kimi K2.7 Code

95.6%#7/86 open

IndependentEpoch AI · Jun 2026

FrontierMath

DeepSeek V4 Pro

45.3%#12/17 open

IndependentEpoch AI · Apr 2026

Kimi K2.7 Code

54.0%#10/17 open

IndependentEpoch AI · Jun 2026

LiveBench Math

DeepSeek V4 Pro

90.7#3/19 open

IndependentLiveBench · Jun 2026

Kimi K2.7 Code

79.6#18/19 open

IndependentLiveBench · Jun 2026

SciCode

DeepSeek V4 Pro

50.0%#12/66 open

Lab-reportedEpoch AI · Apr 2026

Kimi K2.7 Code

47.8%#16/66 open

Lab-reportedEpoch AI · Jun 2026

SimpleQA

DeepSeek V4 Pro

47.0%#4/17 open

IndependentEpoch AI · Apr 2026

Kimi K2.7 Code

36.5%#9/17 open

IndependentEpoch AI · Jun 2026

FrontierMath Tier 4

DeepSeek V4 Pro

2.4%#13/13 open

IndependentEpoch AI · Apr 2026

Kimi K2.7 Code

12.2%#11/13 open

IndependentEpoch AI · Jun 2026

WeirdML

DeepSeek V4 Pro

48.9%#10/46 open

IndependentEpoch AI · Apr 2026

Kimi K2.7 Code

54.1%#8/46 open

IndependentEpoch AI · Jun 2026

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

FrontierCode

DeepSeek V4 Pro

17.6%#9/12 open

Lab-reportedEpoch AI · Apr 2026

Kimi K2.7 Code

30.1%#5/12 open

Lab-reportedEpoch AI · Jun 2026

ALE-Bench

DeepSeek V4 Pro

1006.1#10/43 open

Lab-reportedEpoch AI · Apr 2026

Kimi K2.7 Code

886.2#16/43 open

Lab-reportedEpoch AI · Jun 2026

Chess Puzzles

DeepSeek V4 Pro

19.0%#12/14 open

IndependentEpoch AI · Apr 2026

Kimi K2.7 Code

14.0%#13/14 open

IndependentEpoch AI · Jun 2026

CritPt

DeepSeek V4 Pro

12.9%#9/66 open

Lab-reportedEpoch AI · Apr 2026

Kimi K2.7 Code

10.0%#11/66 open

Lab-reportedEpoch AI · Jun 2026

Surface Evolver Bench

DeepSeek V4 Pro

40.0%#9/13 open

Lab-reportedEpoch AI · Apr 2026

Kimi K2.7 Code

48.8%#5/13 open

Lab-reportedEpoch AI · Jun 2026

Vending-Bench 2

DeepSeek V4 Pro

3284.5#8/17 open

Lab-reportedEpoch AI · Apr 2026

Kimi K2.7 Code

5082.9#6/17 open

Lab-reportedEpoch AI · Jun 2026

LMArena WebDev

DeepSeek V4 Pro

1464.2#20/52 open

ArenaLMArena · Oct 2026

Kimi K2.7 Code

1473.8#19/52 open

ArenaLMArena · Oct 2026

Only DeepSeek V4 Pro (7)

Only Kimi K2.7 Code (3)

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 ProKimi K2.7 Code

Q2_K

DeepSeek V4 Pro

680 GB574.1 GB file

32K 680 GB · 128K 682 GB

Kimi K2.7 Code

440 GB~436.4 GB file

32K 439 GB · 128K 446 GB

Q3_K_M

DeepSeek V4 Pro

780 GB748.7 GB file

32K 780 GB · 128K 782 GB

Kimi K2.7 Code

504 GB463.6 GB file

32K 503 GB · 128K 510 GB

Q4_K_M

DeepSeek V4 Pro

960 GB954.6 GB file

32K 960 GB · 128K 961 GB

Kimi K2.7 Code

620 GB~616.1 GB file

32K 619 GB · 128K 626 GB

Q5_K_M

DeepSeek V4 Pro

1140 GB~1139.2 GB file

32K 1140 GB · 128K 1141 GB

Kimi K2.7 Code

736 GB~731.6 GB file

32K 734 GB · 128K 741 GB

Q6_K

DeepSeek V4 Pro

1319 GB~1319.0 GB file

32K 1320 GB · 128K 1321 GB

Kimi K2.7 Code

851 GB~847.2 GB file

32K 850 GB · 128K 857 GB

Q8_0

DeepSeek V4 Pro

1599 GB1671.8 GB file

32K 1600 GB · 128K 1601 GB

Kimi K2.7 Code

1031 GB~1026.9 GB file

32K 1029 GB · 128K 1036 GB

8 more quants

IQ2_XXS

DeepSeek V4 Pro

not listed

Kimi K2.7 Code

286 GB317.8 GB file

32K 285 GB · 128K 292 GB

IQ2_S

DeepSeek V4 Pro

not listed

Kimi K2.7 Code

325 GB334.8 GB file

32K 324 GB · 128K 330 GB

IQ2_M

DeepSeek V4 Pro

not listed

Kimi K2.7 Code

350 GB318.0 GB file

32K 349 GB · 128K 356 GB

IQ3_XXS

DeepSeek V4 Pro

not listed

Kimi K2.7 Code

402 GB376.9 GB file

32K 401 GB · 128K 407 GB

IQ3_S

DeepSeek V4 Pro

not listed

Kimi K2.7 Code

440 GB418.8 GB file

32K 439 GB · 128K 446 GB

Q3_K_L

DeepSeek V4 Pro

not listed

Kimi K2.7 Code

530 GB493.9 GB file

32K 529 GB · 128K 536 GB

IQ4_XS

DeepSeek V4 Pro

not listed

Kimi K2.7 Code

556 GB495.1 GB file

32K 555 GB · 128K 561 GB

BF16

DeepSeek V4 Pro

3198 GB~3197.7 GB file

32K 3198 GB · 128K 3200 GB

Kimi K2.7 Code

2058 GB~2053.8 GB file

32K 2056 GB · 128K 2063 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 →

Kimi K2.7 Code

Ollama · Q4_K_M

$ollama run hf.co/unsloth/Kimi-K2.7-Code-GGUF

All ways to run Kimi K2.7 Code →

Specifications

Specifications of DeepSeek V4 Pro and Kimi K2.7 Code
Spec	DeepSeek V4 Pro	Kimi K2.7 Code
Publisher	DeepSeek	Moonshot AI
Parameters	1598.8B	1026.9B
Active parameters	51B per token	33B per token
Architecture	DeepseekV4ForCausalLM	KimiK25ForConditionalGeneration
Context	1M tokens	256K tokens
Modalities	Chat	Vision, Code, Function-calling
License	MIT	Other
Released	Apr 2026	Jun 2026

Frequently Asked Questions

Which is better, DeepSeek V4 Pro or Kimi K2.7 Code?

DeepSeek V4 Pro and Kimi K2.7 Code match in llmrun Score, with values of 116.5 versus 118.1; neither fits 24 GB. Choose DeepSeek V4 Pro for longer context window and also does chat; choose Kimi K2.7 Code for lighter on memory and also does vision, code, function-calling.

Which needs less VRAM, DeepSeek V4 Pro or Kimi K2.7 Code?

At Q4_K_M, DeepSeek V4 Pro requires 959.6 GB while Kimi K2.7 Code needs 620.0 GB, making Kimi K2.7 Code the easier model to run on your local setup.

Can I run DeepSeek V4 Pro or Kimi K2.7 Code on a 16 GB GPU?

DeepSeek V4 Pro does not fit (smallest quant needs 680 GB); Kimi K2.7 Code does not fit (smallest quant needs 283.8 GB).

Which has a longer context window, DeepSeek V4 Pro or Kimi K2.7 Code?

DeepSeek V4 Pro supports 1,048,576 tokens and Kimi K2.7 Code supports 262,144 tokens.