Qwen2.5 72B Instruct vs Qwen3.8 27B

Qwen3.8 27B leads llmrun Score 122.1 to 81.7; only Qwen3.8 27B fits 24 GB (Q4_K_M); Qwen2.5 72B Instruct needs 25.7 GB even at its smallest quant.

Model A · Alibaba

Qwen2.5 72B Instruct

81.7llmrun Score

72.7B · 32K context · Other · Sep 2024

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 Qwen2.5 72B Instruct 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 81.7
Stronger at reasoning
124.0 vs 85.6
Fits a 12 GB GPU
IQ3_XXS · 12 GB; Qwen2.5 72B Instruct needs 25.7 GB at its smallest quant
Fits a Mac 32 GB
Q4_K_M · 17.9 GB; Qwen2.5 72B Instruct needs 25.7 GB at its smallest quant
Lighter on memory
17.4 vs 44.6 GB at Q4_K_M
Longer context window
256K vs 32K tokens

On your hardware

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

MemoryQwen2.5 72B InstructQwen3.8 27B

8 GB

RTX 4060 · 8 GB tier

Qwen2.5 72B Instruct

Doesn't fit

smallest quant needs 25.7 GB

Qwen3.8 27B

Doesn't fit

smallest quant needs 8.8 GB

12 GB

RTX 3060 12GB · 12 GB tier

Qwen2.5 72B Instruct

Doesn't fit

smallest quant needs 25.7 GB

Qwen3.8 27B

IQ3_XXS · ~20 tok/s

12 GB · up to 16Kfits

16 GB

RTX 5060 Ti 16GB · 16 GB tier

Qwen2.5 72B Instruct

Doesn't fit

smallest quant needs 25.7 GB

Qwen3.8 27B

Q3_K_L · ~19 tok/s

15.4 GB · up to 16Kfits

24 GBreference

RTX 4090 · 24 GB tier

Qwen2.5 72B Instruct

Doesn't fit

smallest quant needs 25.7 GB

Qwen3.8 27B

Q4_K_M · ~38 tok/s

17.9 GB · up to 64Kfits

32 GB

RTX 5090 · 32 GB tier

Qwen2.5 72B Instruct

IQ2_M · ~46 tok/s

30.2 GB · up to 16K

Qwen3.8 27B

Q4_K_M · ~67 tok/s

17.9 GB · up to 256Kmore headroom

48 GB

RTX A6000 · 48 GB tier

Qwen2.5 72B Instruct

Q4_K_M · ~11 tok/stight

49.3 GB · up to 8K

Qwen3.8 27B

Q4_K_M · ~29 tok/s

17.9 GB · up to 256Kmore headroom

Mac 32 GB

Mac Mini M4 · Macs

Qwen2.5 72B Instruct

IQ2_S · ~4 tok/s

28.4 GB · up to 16K

Qwen3.8 27B

Q4_K_M · ~5 tok/s

17.9 GB · up to 128Kmore headroom

Mac 64 GB

Mac Studio M4 Max · Macs

Qwen2.5 72B Instruct

Q4_K_M · ~9 tok/s

49.3 GB · up to 32K

Qwen3.8 27B

Q4_K_M · ~22 tok/s

17.9 GB · up to 256Kfaster

Mac 128 GB

Mac Studio M4 Max · Macs

Qwen2.5 72B Instruct

Q4_K_M · ~9 tok/s

49.3 GB · up to 32K

Qwen3.8 27B

Q4_K_M · ~22 tok/s

17.9 GB · up to 256Kfaster

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

Qwen2.5 72B Instruct Qwen3.8 27B

Overall

Qwen2.5 72B Instruct

81.777–86

Qwen3.8 27B

122.1120–124

Coding

Qwen2.5 72B Instruct

no data

Qwen3.8 27B

119.7114–125

Agents

Qwen2.5 72B Instruct

no data

Qwen3.8 27B

122.8117–129

Math

Qwen2.5 72B Instruct

no data

Qwen3.8 27B

117.5111–124

Science

Qwen2.5 72B Instruct

no data

Qwen3.8 27B

112.093–131

Reasoning

Qwen2.5 72B Instruct

85.677–94

Qwen3.8 27B

124.0121–127

Qwen2.5 72B Instruct · #80 of 100 open models · #30 of 49 that fit ≤64 GB

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

Benchmarks

2 shared boards: Qwen2.5 72B Instruct leads 0, Qwen3.8 27B leads 2 · Benchmarks imported Oct 2026

BoardQwen2.5 72B InstructQwen3.8 27B

DTBench

Qwen2.5 72B Instruct

62.9%#40/68 open

Lab-reportedEpoch AI · Sep 2024

Qwen3.8 27B

88.0%#9/68 open

Lab-reportedEpoch AI · Aug 2026

LMCA

Qwen2.5 72B Instruct

13.4%#44/54 open

Lab-reportedEpoch AI · Sep 2024

Qwen3.8 27B

41.4%#7/54 open

Lab-reportedEpoch AI · Aug 2026

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

LMArena Text

Qwen2.5 72B Instruct

1302.9#106/187 open

ArenaLMArena · Oct 2026

Qwen3.8 27B

1437.8#22/187 open

ArenaLMArena · Oct 2026

Only Qwen2.5 72B Instruct (9)

Only Qwen3.8 27B (18)

VRAM by quantization

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

QuantQwen2.5 72B InstructQwen3.8 27B

Q2_K

Qwen2.5 72B Instruct

31.9 GB29.8 GB file

32K 41.9 GB · 128K —

Qwen3.8 27B

12.6 GB10.8 GB file

32K 13.9 GB · 128K 19.3 GB

Q3_K_S

Qwen2.5 72B Instruct

32.8 GB34.5 GB file

32K 42.9 GB · 128K —

Qwen3.8 27B

12.9 GB12.7 GB file

32K 14.2 GB · 128K 19.6 GB

Q3_K_M

Qwen2.5 72B Instruct

36.4 GB37.7 GB file

32K 46.5 GB · 128K —

Qwen3.8 27B

14.3 GB13.4 GB file

32K 15.6 GB · 128K 21.0 GB

Q4_0

Qwen2.5 72B Instruct

37.3 GB41.4 GB file

32K 47.4 GB · 128K —

Qwen3.8 27B

14.6 GB~13.9 GB file

32K 16.0 GB · 128K 21.4 GB

Q4_K_M

Qwen2.5 72B Instruct

44.6 GB47.4 GB file

32K 54.7 GB · 128K —

Qwen3.8 27B

17.4 GB16.5 GB file

32K 18.8 GB · 128K 24.1 GB

Q5_K_M

Qwen2.5 72B Instruct

52.8 GB54.4 GB file

32K 62.8 GB · 128K —

Qwen3.8 27B

20.5 GB19.8 GB file

32K 21.9 GB · 128K 27.3 GB

Q6_K

Qwen2.5 72B Instruct

61.0 GB64.3 GB file

32K 71.0 GB · 128K —

Qwen3.8 27B

23.7 GB22.0 GB file

32K 25.0 GB · 128K 30.4 GB

Q8_0

Qwen2.5 72B Instruct

73.7 GB77.3 GB file

32K 83.8 GB · 128K —

Qwen3.8 27B

28.5 GB29.0 GB file

32K 29.9 GB · 128K 35.2 GB

19 more quants

IQ2_XXS

Qwen2.5 72B Instruct

21.0 GB25.5 GB file

32K 31.0 GB · 128K —

Qwen3.8 27B

8.4 GB7.3 GB file

32K 9.7 GB · 128K 15.1 GB

IQ2_XS

Qwen2.5 72B Instruct

22.8 GB27.1 GB file

32K 32.9 GB · 128K —

Qwen3.8 27B

9.1 GB9.1 GB file

32K 10.4 GB · 128K 15.8 GB

IQ2_S

Qwen2.5 72B Instruct

23.7 GB27.9 GB file

32K 33.8 GB · 128K —

Qwen3.8 27B

9.4 GB8.4 GB file

32K 10.8 GB · 128K 16.1 GB

IQ2_M

Qwen2.5 72B Instruct

25.5 GB29.3 GB file

32K 35.6 GB · 128K —

Qwen3.8 27B

10.1 GB10.5 GB file

32K 11.5 GB · 128K 16.8 GB

IQ3_XXS

Qwen2.5 72B Instruct

29.1 GB31.8 GB file

32K 39.2 GB · 128K —

Qwen3.8 27B

11.5 GB10.9 GB file

32K 12.9 GB · 128K 18.2 GB

Q2_K_S

Qwen2.5 72B Instruct

30.1 GB29.6 GB file

32K 40.1 GB · 128K —

Qwen3.8 27B

not listed

IQ3_XS

Qwen2.5 72B Instruct

31.0 GB32.8 GB file

32K 41.0 GB · 128K —

Qwen3.8 27B

12.2 GB10.4 GB file

32K 13.6 GB · 128K 18.9 GB

IQ3_S

Qwen2.5 72B Instruct

31.9 GB34.5 GB file

32K 41.9 GB · 128K —

Qwen3.8 27B

12.6 GB12.0 GB file

32K 13.9 GB · 128K 19.3 GB

IQ3_M

Qwen2.5 72B Instruct

33.7 GB35.5 GB file

32K 43.8 GB · 128K —

Qwen3.8 27B

13.3 GB14.9 GB file

32K 14.6 GB · 128K 20.0 GB

Q3_K_L

Qwen2.5 72B Instruct

38.2 GB39.5 GB file

32K 48.3 GB · 128K —

Qwen3.8 27B

15.0 GB14.1 GB file

32K 16.3 GB · 128K 21.7 GB

IQ4_XS

Qwen2.5 72B Instruct

40.1 GB39.7 GB file

32K 50.1 GB · 128K —

Qwen3.8 27B

15.7 GB14.3 GB file

32K 17.0 GB · 128K 22.4 GB

IQ4_NL

Qwen2.5 72B Instruct

41.9 GB41.3 GB file

32K 51.9 GB · 128K —

Qwen3.8 27B

16.4 GB17.4 GB file

32K 17.7 GB · 128K 23.1 GB

Q4_1

Qwen2.5 72B Instruct

41.9 GB45.7 GB file

32K 51.9 GB · 128K —

Qwen3.8 27B

16.4 GB17.5 GB file

32K 17.7 GB · 128K 23.1 GB

Q4_K_S

Qwen2.5 72B Instruct

41.9 GB43.9 GB file

32K 51.9 GB · 128K —

Qwen3.8 27B

16.4 GB15.4 GB file

32K 17.7 GB · 128K 23.1 GB

Q4_K_L

Qwen2.5 72B Instruct

not listed

Qwen3.8 27B

17.8 GB18.8 GB file

32K 19.1 GB · 128K 24.5 GB

Q5_0

Qwen2.5 72B Instruct

46.4 GB50.3 GB file

32K 56.5 GB · 128K —

Qwen3.8 27B

not listed

Q5_K_S

Qwen2.5 72B Instruct

51.0 GB51.4 GB file

32K 61.0 GB · 128K —

Qwen3.8 27B

19.9 GB18.7 GB file

32K 21.2 GB · 128K 26.6 GB

Q5_K_L

Qwen2.5 72B Instruct

not listed

Qwen3.8 27B

20.9 GB21.5 GB file

32K 22.2 GB · 128K 27.6 GB

BF16

Qwen2.5 72B Instruct

146 GB~145.4 GB file

32K 156 GB · 128K —

Qwen3.8 27B

56.3 GB54.7 GB file

32K 57.7 GB · 128K 63.0 GB

How to run each

Qwen2.5 72B Instruct

Ollama · Q4_K_M

$ollama run hf.co/bartowski/Qwen2.5-72B-Instruct-GGUF:Q4_K_M

All ways to run Qwen2.5 72B Instruct →

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 Qwen2.5 72B Instruct and Qwen3.8 27B
Spec	Qwen2.5 72B Instruct	Qwen3.8 27B
Publisher	Alibaba	Alibaba
Parameters	72.7B	27.8B
Architecture	Qwen2ForCausalLM	Qwen3_5ForConditionalGeneration
Context	32K tokens	256K tokens
Modalities	Chat, Function-calling	Vision, Function-calling
License	Other	Apache 2.0
Released	Sep 2024	Aug 2026

Frequently Asked Questions

Which is better, Qwen2.5 72B Instruct or Qwen3.8 27B?

Qwen3.8 27B leads llmrun Score 122.1 to 81.7; only Qwen3.8 27B fits 24 GB (Q4_K_M); Qwen2.5 72B Instruct needs 25.7 GB even at its smallest quant. Choose Qwen2.5 72B Instruct for also does chat; choose Qwen3.8 27B for higher llmrun Score and stronger at reasoning.

Which needs less VRAM, Qwen2.5 72B Instruct or Qwen3.8 27B?

At Q4_K_M, Qwen2.5 72B Instruct requires 44.6 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 Qwen2.5 72B Instruct or Qwen3.8 27B on a 16 GB GPU?

Qwen2.5 72B Instruct does not fit (smallest quant needs 25.7 GB); Qwen3.8 27B fits at Q3_K_L.

Which has a longer context window, Qwen2.5 72B Instruct or Qwen3.8 27B?

Qwen2.5 72B Instruct supports 32,768 tokens and Qwen3.8 27B supports 262,144 tokens.