Mistral Large 3 675B Instruct 2512 vs Qwen3.8 27B

Qwen3.8 27B leads llmrun Score 122.1 to 89.6; only Qwen3.8 27B fits 24 GB (Q4_K_M); Mistral Large 3 675B Instruct 2512 needs 213.1 GB even at its smallest quant.

Model A · Mistral AI

Mistral Large 3 675B Instruct 2512

89.6llmrun Score

675B · Apache 2.0 · Nov 2025

ChatFunctions

Model B · Alibaba

Qwen3.8 27B

122.1llmrun Score

27.8B · 256K context · Apache 2.0 · Aug 2026

VisionFunctions

Which should you choose?

Choose Mistral Large 3 675B Instruct 2512 if you want…

Also does chat
Qwen3.8 27B doesn't list it

Choose Qwen3.8 27B if you want…

Higher llmrun Score
122.1 vs 89.6
Stronger at reasoning
124.0 vs 88.9
Stronger at coding
119.7 vs 92.7
Ahead on the benchmarks
4 of 4 shared boards
Fits a 12 GB GPU
IQ3_XXS · 12 GB; Mistral Large 3 675B Instruct 2512 needs 213.1 GB at its smallest quant
Fits a Mac 32 GB
Q4_K_M · 17.9 GB; Mistral Large 3 675B Instruct 2512 needs 213.1 GB at its smallest quant

On your hardware

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

MemoryMistral Large 3 675B Instruct 2512Qwen3.8 27B

8 GB

RTX 4060 · 8 GB tier

Mistral Large 3 675B Instruct 2512

Doesn't fit

smallest quant needs 213.1 GB

Qwen3.8 27B

Doesn't fit

smallest quant needs 8.8 GB

12 GB

RTX 3060 12GB · 12 GB tier

Mistral Large 3 675B Instruct 2512

Doesn't fit

smallest quant needs 213.1 GB

Qwen3.8 27B

IQ3_XXS · ~20 tok/s

12 GB · up to 16Kfits

16 GB

RTX 5060 Ti 16GB · 16 GB tier

Mistral Large 3 675B Instruct 2512

Doesn't fit

smallest quant needs 213.1 GB

Qwen3.8 27B

Q3_K_L · ~19 tok/s

15.4 GB · up to 16Kfits

24 GBreference

RTX 4090 · 24 GB tier

Mistral Large 3 675B Instruct 2512

Doesn't fit

smallest quant needs 213.1 GB

Qwen3.8 27B

Q4_K_M · ~38 tok/s

17.9 GB · up to 64Kfits

32 GB

RTX 5090 · 32 GB tier

Mistral Large 3 675B Instruct 2512

Doesn't fit

smallest quant needs 213.1 GB

Qwen3.8 27B

Q4_K_M · ~67 tok/s

17.9 GB · up to 256Kfits

48 GB

RTX A6000 · 48 GB tier

Mistral Large 3 675B Instruct 2512

Doesn't fit

smallest quant needs 213.1 GB

Qwen3.8 27B

Q4_K_M · ~29 tok/s

17.9 GB · up to 256Kfits

Mac 32 GB

Mac Mini M4 · Macs

Mistral Large 3 675B Instruct 2512

Doesn't fit

smallest quant needs 213.1 GB

Qwen3.8 27B

Q4_K_M · ~5 tok/s

17.9 GB · up to 128Kfits

Mac 64 GB

Mac Studio M4 Max · Macs

Mistral Large 3 675B Instruct 2512

Doesn't fit

smallest quant needs 213.1 GB

Qwen3.8 27B

Q4_K_M · ~22 tok/s

17.9 GB · up to 256Kfits

Mac 128 GB

Mac Studio M4 Max · Macs

Mistral Large 3 675B Instruct 2512

Doesn't fit

smallest quant needs 213.1 GB

Qwen3.8 27B

Q4_K_M · ~22 tok/s

17.9 GB · up to 256Kfits

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

Mistral Large 3 675B Instruct 2512 Qwen3.8 27B

Overall

Mistral Large 3 675B Instruct 2512

89.682–97

Qwen3.8 27B

122.1120–124

Coding

Mistral Large 3 675B Instruct 2512

92.776–109

Qwen3.8 27B

119.7114–125

Agents

Mistral Large 3 675B Instruct 2512

no data

Qwen3.8 27B

122.8117–129

Math

Mistral Large 3 675B Instruct 2512

no data

Qwen3.8 27B

117.5111–124

Science

Mistral Large 3 675B Instruct 2512

96.875–119

Qwen3.8 27B

112.093–131

Reasoning

Mistral Large 3 675B Instruct 2512

88.980–98

Qwen3.8 27B

124.0121–127

Mistral Large 3 675B Instruct 2512 · #69 of 100 open models · #66 of 96 that fit ≤512 GB

Qwen3.8 27B · #11 of 100 open models · #1 of 35 that fit 24 GB

Benchmarks

4 shared boards: Mistral Large 3 675B Instruct 2512 leads 0, Qwen3.8 27B leads 4 · Benchmarks imported Oct 2026

BoardMistral Large 3 675B Instruct 2512Qwen3.8 27B

SciCode

Mistral Large 3 675B Instruct 2512

36.6%#39/66 open

Lab-reportedEpoch AI · Dec 2025

Qwen3.8 27B

46.6%#19/66 open

Lab-reportedEpoch AI · Aug 2026

CritPt

Mistral Large 3 675B Instruct 2512

0.0%#59/66 open

Lab-reportedEpoch AI · Dec 2025

Qwen3.8 27B

5.4%#15/66 open

Lab-reportedEpoch AI · Aug 2026

DTBench

Mistral Large 3 675B Instruct 2512

65.1%#37/68 open

Lab-reportedEpoch AI · Dec 2025

Qwen3.8 27B

88.0%#9/68 open

Lab-reportedEpoch AI · Aug 2026

LMCA

Mistral Large 3 675B Instruct 2512

16.7%#38/54 open

Lab-reportedEpoch AI · Dec 2025

Qwen3.8 27B

41.4%#7/54 open

Lab-reportedEpoch AI · Aug 2026

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

LMArena Text

Mistral Large 3 675B Instruct 2512

1414.2#41/187 open

ArenaLMArena · Oct 2026

Qwen3.8 27B

1437.8#22/187 open

ArenaLMArena · Oct 2026

LMArena WebDev

Mistral Large 3 675B Instruct 2512

1230.5#50/52 open

ArenaLMArena · Oct 2026

Qwen3.8 27B

1593.2#10/52 open

ArenaLMArena · Oct 2026

Only Mistral Large 3 675B Instruct 2512 (1)

ALE-Bench264.7

Only Qwen3.8 27B (15)

VRAM by quantization

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

QuantMistral Large 3 675B Instruct 2512Qwen3.8 27B

Q2_K

Mistral Large 3 675B Instruct 2512

316 GB247.1 GB file

32K 335 GB · 128K 396 GB

Qwen3.8 27B

12.6 GB10.8 GB file

32K 13.9 GB · 128K 19.3 GB

Q3_K_S

Mistral Large 3 675B Instruct 2512

325 GB291.9 GB file

32K 344 GB · 128K 405 GB

Qwen3.8 27B

12.9 GB12.7 GB file

32K 14.2 GB · 128K 19.6 GB

Q3_K_M

Mistral Large 3 675B Instruct 2512

362 GB321.8 GB file

32K 381 GB · 128K 442 GB

Qwen3.8 27B

14.3 GB13.4 GB file

32K 15.6 GB · 128K 21.0 GB

Q4_0

Mistral Large 3 675B Instruct 2512

371 GB381.7 GB file

32K 390 GB · 128K 451 GB

Qwen3.8 27B

14.6 GB~13.9 GB file

32K 16.0 GB · 128K 21.4 GB

Q4_K_M

Mistral Large 3 675B Instruct 2512

446 GB407.0 GB file

32K 465 GB · 128K 525 GB

Qwen3.8 27B

17.4 GB16.5 GB file

32K 18.8 GB · 128K 24.1 GB

Q5_K_M

Mistral Large 3 675B Instruct 2512

529 GB478.0 GB file

32K 548 GB · 128K 609 GB

Qwen3.8 27B

20.5 GB19.8 GB file

32K 21.9 GB · 128K 27.3 GB

Q6_K

Mistral Large 3 675B Instruct 2512

613 GB553.3 GB file

32K 632 GB · 128K 693 GB

Qwen3.8 27B

23.7 GB22.0 GB file

32K 25.0 GB · 128K 30.4 GB

Q8_0

Mistral Large 3 675B Instruct 2512

743 GB715.7 GB file

32K 762 GB · 128K 822 GB

Qwen3.8 27B

28.5 GB29.0 GB file

32K 29.9 GB · 128K 35.2 GB

17 more quants

IQ2_XXS

Mistral Large 3 675B Instruct 2512

204 GB224.8 GB file

32K 223 GB · 128K 284 GB

Qwen3.8 27B

8.4 GB7.3 GB file

32K 9.7 GB · 128K 15.1 GB

IQ2_XS

Mistral Large 3 675B Instruct 2512

223 GB189.5 GB file

32K 242 GB · 128K 303 GB

Qwen3.8 27B

9.1 GB9.1 GB file

32K 10.4 GB · 128K 15.8 GB

IQ2_S

Mistral Large 3 675B Instruct 2512

232 GB190.8 GB file

32K 251 GB · 128K 312 GB

Qwen3.8 27B

9.4 GB8.4 GB file

32K 10.8 GB · 128K 16.1 GB

IQ2_M

Mistral Large 3 675B Instruct 2512

251 GB235.4 GB file

32K 270 GB · 128K 331 GB

Qwen3.8 27B

10.1 GB10.5 GB file

32K 11.5 GB · 128K 16.8 GB

IQ3_XXS

Mistral Large 3 675B Instruct 2512

288 GB279.2 GB file

32K 307 GB · 128K 368 GB

Qwen3.8 27B

11.5 GB10.9 GB file

32K 12.9 GB · 128K 18.2 GB

IQ3_XS

Mistral Large 3 675B Instruct 2512

306 GB279.1 GB file

32K 325 GB · 128K 386 GB

Qwen3.8 27B

12.2 GB10.4 GB file

32K 13.6 GB · 128K 18.9 GB

IQ3_S

Mistral Large 3 675B Instruct 2512

not listed

Qwen3.8 27B

12.6 GB12.0 GB file

32K 13.9 GB · 128K 19.3 GB

IQ3_M

Mistral Large 3 675B Instruct 2512

334 GB309.9 GB file

32K 353 GB · 128K 414 GB

Qwen3.8 27B

13.3 GB14.9 GB file

32K 14.6 GB · 128K 20.0 GB

Q3_K_L

Mistral Large 3 675B Instruct 2512

381 GB321.7 GB file

32K 400 GB · 128K 460 GB

Qwen3.8 27B

15.0 GB14.1 GB file

32K 16.3 GB · 128K 21.7 GB

IQ4_XS

Mistral Large 3 675B Instruct 2512

399 GB360.0 GB file

32K 418 GB · 128K 479 GB

Qwen3.8 27B

15.7 GB14.3 GB file

32K 17.0 GB · 128K 22.4 GB

IQ4_NL

Mistral Large 3 675B Instruct 2512

418 GB380.7 GB file

32K 437 GB · 128K 498 GB

Qwen3.8 27B

16.4 GB17.4 GB file

32K 17.7 GB · 128K 23.1 GB

Q4_1

Mistral Large 3 675B Instruct 2512

418 GB422.4 GB file

32K 437 GB · 128K 498 GB

Qwen3.8 27B

16.4 GB17.5 GB file

32K 17.7 GB · 128K 23.1 GB

Q4_K_S

Mistral Large 3 675B Instruct 2512

418 GB382.7 GB file

32K 437 GB · 128K 498 GB

Qwen3.8 27B

16.4 GB15.4 GB file

32K 17.7 GB · 128K 23.1 GB

Q4_K_L

Mistral Large 3 675B Instruct 2512

not listed

Qwen3.8 27B

17.8 GB18.8 GB file

32K 19.1 GB · 128K 24.5 GB

Q5_K_S

Mistral Large 3 675B Instruct 2512

510 GB464.4 GB file

32K 530 GB · 128K 590 GB

Qwen3.8 27B

19.9 GB18.7 GB file

32K 21.2 GB · 128K 26.6 GB

Q5_K_L

Mistral Large 3 675B Instruct 2512

not listed

Qwen3.8 27B

20.9 GB21.5 GB file

32K 22.2 GB · 128K 27.6 GB

BF16

Mistral Large 3 675B Instruct 2512

1485 GB1347.0 GB file

32K 1504 GB · 128K 1565 GB

Qwen3.8 27B

56.3 GB54.7 GB file

32K 57.7 GB · 128K 63.0 GB

How to run each

Mistral Large 3 675B Instruct 2512

Ollama · Q4_K_M

$ollama run hf.co/unsloth/Mistral-Large-3-675B-Instruct-2512-GGUF:Q4_K_M

All ways to run Mistral Large 3 675B Instruct 2512 →

Qwen3.8 27B

Ollama · Q4_K_M

$ollama run hf.co/unsloth/Qwen3.8-27B-GGUF:Q4_K_M

All ways to run Qwen3.8 27B →

Specifications

Specifications of Mistral Large 3 675B Instruct 2512 and Qwen3.8 27B
Spec	Mistral Large 3 675B Instruct 2512	Qwen3.8 27B
Publisher	Mistral AI	Alibaba
Parameters	675B	27.8B
Architecture	—	Qwen3_5ForConditionalGeneration
Context	—	256K tokens
Modalities	Chat, Function-calling	Vision, Function-calling
License	Apache 2.0	Apache 2.0
Released	Nov 2025	Aug 2026

Frequently Asked Questions

Which is better, Mistral Large 3 675B Instruct 2512 or Qwen3.8 27B?

Qwen3.8 27B leads llmrun Score 122.1 to 89.6; only Qwen3.8 27B fits 24 GB (Q4_K_M); Mistral Large 3 675B Instruct 2512 needs 213.1 GB even at its smallest quant. Choose Mistral Large 3 675B Instruct 2512 for also does chat; choose Qwen3.8 27B for higher llmrun Score and stronger at reasoning.

Which needs less VRAM, Mistral Large 3 675B Instruct 2512 or Qwen3.8 27B?

At Q4_K_M, Mistral Large 3 675B Instruct 2512 requires 445.5 GB while Qwen3.8 27B needs 17.4 GB, making Qwen3.8 27B the easier model to run on your local setup.

Can I run Mistral Large 3 675B Instruct 2512 or Qwen3.8 27B on a 16 GB GPU?

Mistral Large 3 675B Instruct 2512 does not fit (smallest quant needs 213.1 GB); Qwen3.8 27B fits at Q3_K_L.