DeepSeek v3 vs DeepSeek v3 0324

DeepSeek v3 0324 leads llmrun Score 93.9 to 88.7 (ranges overlap); neither fits 24 GB.

Model A · DeepSeek

DeepSeek v3

88.7llmrun Score

684.5B · 40.1B active · 160K context · Dec 2024

ChatFunctions

Model B · DeepSeek

DeepSeek v3 0324

93.9llmrun Score

684.5B · 40.1B active · 160K context · MIT · Mar 2025

ChatFunctions

Which should you choose?

Choose DeepSeek v3 0324 if you want…

Higher llmrun Score
93.9 vs 88.7, ranges overlap
Stronger at science
97.9 vs 89.5
Stronger at reasoning
87.7 vs 81.1
Ahead on the benchmarks
7 of 8 shared boards

DeepSeek v3: no clear edge over DeepSeek v3 0324 in the data we have.

On your hardware

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

MemoryDeepSeek v3DeepSeek v3 0324

8 GB

RTX 4060 · 8 GB tier

DeepSeek v3

Doesn't fit

smallest quant needs 189.7 GB

DeepSeek v3 0324

Doesn't fit

smallest quant needs 189.7 GB

12 GB

RTX 3060 12GB · 12 GB tier

DeepSeek v3

Doesn't fit

smallest quant needs 189.7 GB

DeepSeek v3 0324

Doesn't fit

smallest quant needs 189.7 GB

16 GB

RTX 5060 Ti 16GB · 16 GB tier

DeepSeek v3

Doesn't fit

smallest quant needs 189.7 GB

DeepSeek v3 0324

Doesn't fit

smallest quant needs 189.7 GB

24 GBreference

RTX 4090 · 24 GB tier

DeepSeek v3

Doesn't fit

smallest quant needs 189.7 GB

DeepSeek v3 0324

Doesn't fit

smallest quant needs 189.7 GB

32 GB

RTX 5090 · 32 GB tier

DeepSeek v3

Doesn't fit

smallest quant needs 189.7 GB

DeepSeek v3 0324

Doesn't fit

smallest quant needs 189.7 GB

48 GB

RTX A6000 · 48 GB tier

DeepSeek v3

Doesn't fit

smallest quant needs 189.7 GB

DeepSeek v3 0324

Doesn't fit

smallest quant needs 189.7 GB

Mac 32 GB

Mac Mini M4 · Macs

DeepSeek v3

Doesn't fit

smallest quant needs 189.7 GB

DeepSeek v3 0324

Doesn't fit

smallest quant needs 189.7 GB

Mac 64 GB

Mac Studio M4 Max · Macs

DeepSeek v3

Doesn't fit

smallest quant needs 189.7 GB

DeepSeek v3 0324

Doesn't fit

smallest quant needs 189.7 GB

Mac 128 GB

Mac Studio M4 Max · Macs

DeepSeek v3

Doesn't fit

smallest quant needs 189.7 GB

DeepSeek v3 0324

Doesn't fit

smallest quant needs 189.7 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 v3 DeepSeek v3 0324

Overall

DeepSeek v3

88.785–92

DeepSeek v3 0324

93.991–97

Coding

DeepSeek v3

96.991–103

DeepSeek v3 0324

99.594–105

Science

DeepSeek v3

89.582–97

DeepSeek v3 0324

97.991–105

Reasoning

DeepSeek v3

81.165–97

DeepSeek v3 0324

87.780–96

DeepSeek v3 · #71 of 100 open models · #67 of 96 that fit ≤512 GB

DeepSeek v3 0324 · #66 of 100 open models · #63 of 96 that fit ≤512 GB

Benchmarks

8 shared boards: DeepSeek v3 leads 0, DeepSeek v3 0324 leads 7 · Benchmarks imported Oct 2026

BoardDeepSeek v3DeepSeek v3 0324

GPQA Diamond

DeepSeek v3

56.5%#45/98 open

IndependentEpoch AI · Dec 2024

DeepSeek v3 0324

67.6%#35/98 open

IndependentEpoch AI · Mar 2025

AIME 2024/2025

DeepSeek v3

15.8%#56/86 open

IndependentEpoch AI · Dec 2024

DeepSeek v3 0324

37.8%#48/86 open

IndependentEpoch AI · Mar 2025

MMLU-Pro

DeepSeek v3

75.9%#29/145 open

IndependentTIGER-Lab

DeepSeek v3 0324

81.3%#20/145 open

IndependentTIGER-Lab

Aider Polyglot

DeepSeek v3

48.4%#9/18 open

IndependentEpoch AI · Dec 2024

DeepSeek v3 0324

55.1%#8/18 open

IndependentEpoch AI · Mar 2025

SciCode

DeepSeek v3

35.8%#44/66 open

Lab-reportedEpoch AI · Dec 2024

DeepSeek v3 0324

39.0%#32/66 open

Lab-reportedEpoch AI · Mar 2025

SimpleBench

DeepSeek v3

18.9%#24/25 open

IndependentEpoch AI · Dec 2024

DeepSeek v3 0324

27.2%#18/25 open

IndependentEpoch AI · Mar 2025

MATH Level 5

DeepSeek v3

64.8%#10/39 open

IndependentEpoch AI · Dec 2024

DeepSeek v3 0324

75.5%#5/39 open

IndependentEpoch AI · Mar 2025

CritPt

DeepSeek v3

0.0%#47/66 open

Lab-reportedEpoch AI · Dec 2024

DeepSeek v3 0324

0.0%#48/66 open

Lab-reportedEpoch AI · Mar 2025

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

LMArena Text

DeepSeek v3

1358.4#70/187 open

ArenaLMArena · Oct 2026

DeepSeek v3 0324

1395.2#52/187 open

ArenaLMArena · Oct 2026

METR Time Horizons

DeepSeek v3

47.4%#6/9 open

Lab-reportedEpoch AI · Dec 2024

DeepSeek v3 0324

49.6%#5/9 open

Lab-reportedEpoch AI · Mar 2025

Only DeepSeek v3 (8)

Only DeepSeek v3 0324 (5)

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 v3DeepSeek v3 0324

Q2_K

DeepSeek v3

295 GB~290.9 GB file

32K 294 GB · 128K 300 GB

DeepSeek v3 0324

295 GB244.0 GB file

32K 294 GB · 128K 300 GB

Q3_K_S

DeepSeek v3

not listed

DeepSeek v3 0324

303 GB289.1 GB file

32K 302 GB · 128K 309 GB

Q3_K_M

DeepSeek v3

338 GB319.2 GB file

32K 336 GB · 128K 343 GB

DeepSeek v3 0324

338 GB319.2 GB file

32K 336 GB · 128K 343 GB

Q4_0

DeepSeek v3

not listed

DeepSeek v3 0324

346 GB379.0 GB file

32K 345 GB · 128K 352 GB

Q4_K_M

DeepSeek v3

415 GB404.4 GB file

32K 413 GB · 128K 420 GB

DeepSeek v3 0324

415 GB404.4 GB file

32K 413 GB · 128K 420 GB

Q5_K_M

DeepSeek v3

492 GB475.4 GB file

32K 490 GB · 128K 497 GB

DeepSeek v3 0324

492 GB475.4 GB file

32K 490 GB · 128K 497 GB

Q6_K

DeepSeek v3

569 GB550.8 GB file

32K 567 GB · 128K 574 GB

DeepSeek v3 0324

569 GB550.8 GB file

32K 567 GB · 128K 574 GB

Q8_0

DeepSeek v3

688 GB713.3 GB file

32K 687 GB · 128K 694 GB

DeepSeek v3 0324

688 GB713.3 GB file

32K 687 GB · 128K 694 GB

15 more quants

IQ2_XXS

DeepSeek v3

192 GB174.4 GB file

32K 191 GB · 128K 198 GB

DeepSeek v3 0324

192 GB218.7 GB file

32K 191 GB · 128K 198 GB

IQ2_XS

DeepSeek v3

209 GB195.1 GB file

32K 208 GB · 128K 215 GB

DeepSeek v3 0324

not listed

IQ2_S

DeepSeek v3

218 GB197.0 GB file

32K 217 GB · 128K 223 GB

DeepSeek v3 0324

218 GB197.0 GB file

32K 217 GB · 128K 223 GB

IQ2_M

DeepSeek v3

235 GB217.4 GB file

32K 234 GB · 128K 241 GB

DeepSeek v3 0324

235 GB217.4 GB file

32K 234 GB · 128K 241 GB

IQ3_XXS

DeepSeek v3

269 GB257.9 GB file

32K 268 GB · 128K 275 GB

DeepSeek v3 0324

269 GB257.9 GB file

32K 268 GB · 128K 275 GB

IQ3_XS

DeepSeek v3

286 GB272.8 GB file

32K 285 GB · 128K 292 GB

DeepSeek v3 0324

not listed

IQ3_S

DeepSeek v3

295 GB289.1 GB file

32K 294 GB · 128K 300 GB

DeepSeek v3 0324

not listed

IQ3_M

DeepSeek v3

312 GB292.1 GB file

32K 311 GB · 128K 318 GB

DeepSeek v3 0324

not listed

Q3_K_L

DeepSeek v3

not listed

DeepSeek v3 0324

355 GB347.4 GB file

32K 353 GB · 128K 360 GB

IQ4_XS

DeepSeek v3

not listed

DeepSeek v3 0324

372 GB357.1 GB file

32K 371 GB · 128K 377 GB

IQ4_NL

DeepSeek v3

389 GB378.1 GB file

32K 388 GB · 128K 395 GB

DeepSeek v3 0324

389 GB378.1 GB file

32K 388 GB · 128K 395 GB

Q4_1

DeepSeek v3

not listed

DeepSeek v3 0324

389 GB419.9 GB file

32K 388 GB · 128K 395 GB

Q4_K_S

DeepSeek v3

not listed

DeepSeek v3 0324

389 GB380.0 GB file

32K 388 GB · 128K 395 GB

Q5_K_S

DeepSeek v3

not listed

DeepSeek v3 0324

475 GB461.8 GB file

32K 473 GB · 128K 480 GB

BF16

DeepSeek v3

1373 GB1342.3 GB file

32K 1372 GB · 128K 1379 GB

DeepSeek v3 0324

1373 GB1342.3 GB file

32K 1372 GB · 128K 1379 GB

How to run each

DeepSeek v3

Ollama · Q4_K_M

$ollama run hf.co/unsloth/DeepSeek-V3-GGUF:Q4_K_M

All ways to run DeepSeek v3 →

DeepSeek v3 0324

Ollama · Q4_K_M

$ollama run hf.co/unsloth/DeepSeek-V3-0324-GGUF:Q4_K_M

All ways to run DeepSeek v3 0324 →

Specifications

Specifications of DeepSeek v3 and DeepSeek v3 0324
Spec	DeepSeek v3	DeepSeek v3 0324
Publisher	DeepSeek	DeepSeek
Parameters	684.5B	684.5B
Active parameters	40B per token	40B per token
Architecture	DeepseekV3ForCausalLM	DeepseekV3ForCausalLM
Context	160K tokens	160K tokens
Modalities	Chat, Function-calling	Chat, Function-calling
License	—	MIT
Released	Dec 2024	Mar 2025

Frequently Asked Questions

Which is better, DeepSeek v3 or DeepSeek v3 0324?

DeepSeek v3 0324 leads llmrun Score 93.9 to 88.7 (ranges overlap); neither fits 24 GB. Choose DeepSeek v3 0324 for higher llmrun Score and stronger at science.

Which needs less VRAM, DeepSeek v3 or DeepSeek v3 0324?

At Q4_K_M, DeepSeek v3 requires 414.6 GB while DeepSeek v3 0324 needs 414.6 GB, making DeepSeek v3 the easier model to run on your local setup.

Can I run DeepSeek v3 or DeepSeek v3 0324 on a 16 GB GPU?

DeepSeek v3 does not fit (smallest quant needs 189.7 GB); DeepSeek v3 0324 does not fit (smallest quant needs 189.7 GB).

Which has a longer context window, DeepSeek v3 or DeepSeek v3 0324?

DeepSeek v3 supports 163,840 tokens and DeepSeek v3 0324 supports 163,840 tokens.