llmrun

Qwen3 14B vs Qwen3 30B A3B Instruct 2507

Qwen3 14B and Qwen3 30B A3B Instruct 2507 match in llmrun Score, with values of 97.9 versus 96.5; both fit 24 GB at Q4_K_M; Qwen3 30B A3B Instruct 2507 is 2.5× faster on RTX 4090 (~173 vs ~69 tok/s).

Model A · Alibaba

Qwen3 14B
97.9llmrun Score

14.8B · 40K context · Apache 2.0 · Apr 2025

ChatFunctions

Model B · Alibaba

Qwen3 30B A3B Instruct 2507
96.5llmrun Score

30.5B · 3.4B active · 256K context · Apache 2.0 · Jul 2025

ChatFunctions

Which should you choose?

Choose Qwen3 14B if you want…

  • Fits a 12 GB GPU

    Q4_K_M · 11.9 GB; Qwen3 30B A3B Instruct 2507 needs 9.5 GB at its smallest quant

  • Lighter on memory

    9.5 vs 18.7 GB at Q4_K_M

Choose Qwen3 30B A3B Instruct 2507 if you want…

  • 2.5× faster on RTX 4090

    ~173 vs ~69 tok/s

  • Longer context window

    256K vs 40K tokens

  • Newer release

    Jul 2025 vs Apr 2025

On your hardware

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

8 GB

RTX 4060 · 8 GB tier

Qwen3 14B

IQ2_M · ~32 tok/s

8 GB · up to 16Kfits

Qwen3 30B A3B Instruct 2507

Doesn't fit

smallest quant needs 9.5 GB

12 GB

RTX 3060 12GB · 12 GB tier

Qwen3 14B

Q4_K_M · ~25 tok/s

11.9 GB · up to 16Kmore headroom

Qwen3 30B A3B Instruct 2507

IQ2_M · ~130 tok/s

11.4 GB · up to 16K

16 GB

RTX 5060 Ti 16GB · 16 GB tier

Qwen3 14B

Q4_K_M · ~31 tok/s

11.9 GB · up to 40Kmore headroom

Qwen3 30B A3B Instruct 2507

Q3_K_M · ~119 tok/s

16 GB · up to 16K

24 GBreference

RTX 4090 · 24 GB tier

Qwen3 14B

Q4_K_M · ~69 tok/s

11.9 GB · up to 40K

Qwen3 30B A3B Instruct 2507

Q4_K_M · ~173 tok/s

19.4 GB · up to 64Kfaster

32 GB

RTX 5090 · 32 GB tier

Qwen3 14B

Q4_K_M · ~123 tok/s

11.9 GB · up to 40K

Qwen3 30B A3B Instruct 2507

Q4_K_M · ~227 tok/s

19.4 GB · up to 256Kfaster

48 GB

RTX A6000 · 48 GB tier

Qwen3 14B

Q4_K_M · ~53 tok/s

11.9 GB · up to 40K

Qwen3 30B A3B Instruct 2507

Q4_K_M · ~148 tok/s

19.4 GB · up to 256Kfaster

Mac 32 GB

Mac Mini M4 · Macs

Qwen3 14B

Q4_K_M · ~9 tok/s

11.9 GB · up to 40K

Qwen3 30B A3B Instruct 2507

Q4_K_M · ~29 tok/s

19.4 GB · up to 128Kfaster

Mac 64 GB

Mac Studio M4 Max · Macs

Qwen3 14B

Q4_K_M · ~40 tok/s

11.9 GB · up to 40K

Qwen3 30B A3B Instruct 2507

Q4_K_M · ~67 tok/s

19.4 GB · up to 256Kfaster

Mac 128 GB

Mac Studio M4 Max · Macs

Qwen3 14B

Q4_K_M · ~40 tok/s

11.9 GB · up to 40K

Qwen3 30B A3B Instruct 2507

Q4_K_M · ~67 tok/s

19.4 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

Overall

Qwen3 14B

97.995–101

Qwen3 30B A3B Instruct 2507

96.593–100

Science

Qwen3 14B

94.187–101

Qwen3 30B A3B Instruct 2507

no data

Reasoning

Qwen3 14B

87.879–97

Qwen3 30B A3B Instruct 2507

91.984–100

Qwen3 14B · #53 of 92 open models · #12 of 34 that fit 24 GB

Qwen3 30B A3B Instruct 2507 · #56 of 92 open models · #14 of 34 that fit 24 GB

Benchmarks

5 shared boards: Qwen3 14B leads 3, Qwen3 30B A3B Instruct 2507 leads 2 · Benchmarks imported Oct 2026

GPQA Diamond

Qwen3 14B

63.8%#39/97 open

IndependentEpoch AI · Apr 2025

Qwen3 30B A3B Instruct 2507

55.6%#48/97 open

IndependentEpoch AI · Jul 2025

AIME 2024/2025

Qwen3 14B

66.4%#32/85 open

IndependentEpoch AI · Apr 2025

Qwen3 30B A3B Instruct 2507

62.2%#35/85 open

IndependentEpoch AI · Jul 2025

Chess Puzzles

Qwen3 14B

4.0%#34/67 open

IndependentEpoch AI · Apr 2025

Qwen3 30B A3B Instruct 2507

2.0%#39/67 open

IndependentEpoch AI · Jul 2025

DTBench

Qwen3 14B

64.0%#39/68 open

Lab-reportedEpoch AI · Apr 2025

Qwen3 30B A3B Instruct 2507

67.2%#36/68 open

Lab-reportedEpoch AI · Jul 2025

LMCA

Qwen3 14B

18.2%#35/54 open

Lab-reportedEpoch AI · Apr 2025

Qwen3 30B A3B Instruct 2507

22.4%#31/54 open

Lab-reportedEpoch AI · Jul 2025

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

Only Qwen3 14B (3)

Only Qwen3 30B A3B Instruct 2507 (1)

VRAM by quantization

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

Q2_K

Qwen3 14B

6.9 GB5.8 GB file

32K 11.9 GB · 128K —

Qwen3 30B A3B Instruct 2507

13.4 GB11.3 GB file

32K 14.9 GB · 128K 19.7 GB

Q3_K_S

Qwen3 14B

7.1 GB6.7 GB file

32K 12.1 GB · 128K —

Qwen3 30B A3B Instruct 2507

13.8 GB13.3 GB file

32K 15.3 GB · 128K 20.1 GB

Q3_K_M

Qwen3 14B

7.8 GB7.3 GB file

32K 12.9 GB · 128K —

Qwen3 30B A3B Instruct 2507

15.3 GB14.7 GB file

32K 16.8 GB · 128K 21.6 GB

Q4_0

Qwen3 14B

8.0 GB8.5 GB file

32K 13.1 GB · 128K —

Qwen3 30B A3B Instruct 2507

15.7 GB17.4 GB file

32K 17.2 GB · 128K 22.0 GB

Q4_K_M

Qwen3 14B

9.5 GB9.0 GB file

32K 14.5 GB · 128K —

Qwen3 30B A3B Instruct 2507

18.7 GB18.6 GB file

32K 20.2 GB · 128K 25.1 GB

Q5_K_M

Qwen3 14B

11.2 GB10.5 GB file

32K 16.2 GB · 128K —

Qwen3 30B A3B Instruct 2507

22.2 GB21.7 GB file

32K 23.7 GB · 128K 28.5 GB

Q6_K

Qwen3 14B

12.8 GB12.1 GB file

32K 17.9 GB · 128K —

Qwen3 30B A3B Instruct 2507

25.6 GB25.1 GB file

32K 27.1 GB · 128K 31.9 GB

Q8_0

Qwen3 14B

15.4 GB15.7 GB file

32K 20.4 GB · 128K —

Qwen3 30B A3B Instruct 2507

30.9 GB32.5 GB file

32K 32.4 GB · 128K 37.3 GB

19 more quants

IQ2_XXS

Qwen3 14B

4.7 GB4.5 GB file

32K 9.7 GB · 128K —

Qwen3 30B A3B Instruct 2507

8.8 GB10.3 GB file

32K 10.3 GB · 128K 15.1 GB

IQ2_XS

Qwen3 14B

5.1 GB4.7 GB file

32K 10.1 GB · 128K —

Qwen3 30B A3B Instruct 2507

9.6 GB8.7 GB file

32K 11.1 GB · 128K 15.9 GB

IQ2_S

Qwen3 14B

5.3 GB5.0 GB file

32K 10.3 GB · 128K —

Qwen3 30B A3B Instruct 2507

9.9 GB8.7 GB file

32K 11.5 GB · 128K 16.3 GB

IQ2_M

Qwen3 14B

5.6 GB5.4 GB file

32K 10.7 GB · 128K —

Qwen3 30B A3B Instruct 2507

10.7 GB10.8 GB file

32K 12.2 GB · 128K 17.1 GB

IQ3_XXS

Qwen3 14B

6.4 GB6.0 GB file

32K 11.4 GB · 128K —

Qwen3 30B A3B Instruct 2507

12.2 GB12.9 GB file

32K 13.7 GB · 128K 18.6 GB

Q2_K_S

Qwen3 14B

6.5 GB5.4 GB file

32K 11.6 GB · 128K —

Qwen3 30B A3B Instruct 2507

12.6 GB10.5 GB file

32K 14.1 GB · 128K 19.0 GB

IQ3_XS

Qwen3 14B

6.7 GB6.4 GB file

32K 11.8 GB · 128K —

Qwen3 30B A3B Instruct 2507

13.0 GB12.7 GB file

32K 14.5 GB · 128K 19.3 GB

IQ3_S

Qwen3 14B

6.9 GB6.7 GB file

32K 11.9 GB · 128K —

Qwen3 30B A3B Instruct 2507

13.4 GB10.3 GB file

32K 14.9 GB · 128K 19.7 GB

IQ3_M

Qwen3 14B

7.3 GB6.9 GB file

32K 12.3 GB · 128K —

Qwen3 30B A3B Instruct 2507

14.1 GB14.1 GB file

32K 15.7 GB · 128K 20.5 GB

Q3_K_L

Qwen3 14B

8.2 GB7.9 GB file

32K 13.2 GB · 128K —

Qwen3 30B A3B Instruct 2507

16.1 GB15.9 GB file

32K 17.6 GB · 128K 22.4 GB

IQ4_XS

Qwen3 14B

8.6 GB8.1 GB file

32K 13.6 GB · 128K —

Qwen3 30B A3B Instruct 2507

16.8 GB16.4 GB file

32K 18.3 GB · 128K 23.2 GB

IQ4_NL

Qwen3 14B

8.9 GB8.5 GB file

32K 14.0 GB · 128K —

Qwen3 30B A3B Instruct 2507

17.6 GB17.3 GB file

32K 19.1 GB · 128K 23.9 GB

Q4_1

Qwen3 14B

8.9 GB9.4 GB file

32K 14.0 GB · 128K —

Qwen3 30B A3B Instruct 2507

17.6 GB19.2 GB file

32K 19.1 GB · 128K 23.9 GB

Q4_K_S

Qwen3 14B

8.9 GB8.6 GB file

32K 14.0 GB · 128K —

Qwen3 30B A3B Instruct 2507

17.6 GB17.5 GB file

32K 19.1 GB · 128K 23.9 GB

Q4_K_L

Qwen3 14B

9.7 GB9.6 GB file

32K 14.7 GB · 128K —

Qwen3 30B A3B Instruct 2507

19.1 GB18.9 GB file

32K 20.6 GB · 128K 25.4 GB

Q5_0

Qwen3 14B

9.9 GB10.3 GB file

32K 14.9 GB · 128K —

Qwen3 30B A3B Instruct 2507

19.5 GB21.1 GB file

32K 21.0 GB · 128K 25.8 GB

Q5_K_S

Qwen3 14B

10.8 GB10.3 GB file

32K 15.8 GB · 128K —

Qwen3 30B A3B Instruct 2507

21.4 GB21.1 GB file

32K 22.9 GB · 128K 27.7 GB

Q5_K_L

Qwen3 14B

11.3 GB11.0 GB file

32K 16.4 GB · 128K —

Qwen3 30B A3B Instruct 2507

22.5 GB21.9 GB file

32K 24.1 GB · 128K 28.9 GB

BF16

Qwen3 14B

30.2 GB29.5 GB file

32K 35.2 GB · 128K —

Qwen3 30B A3B Instruct 2507

61.5 GB61.1 GB file

32K 63.0 GB · 128K 67.8 GB

How to run each

Specifications

Specifications of Qwen3 14B and Qwen3 30B A3B Instruct 2507
SpecQwen3 14BQwen3 30B A3B Instruct 2507
PublisherAlibabaAlibaba
Parameters14.8B30.5B
Active parameters—3.4B per token
ArchitectureQwen3ForCausalLMQwen3MoeForCausalLM
Context40K tokens256K tokens
ModalitiesChat, Function-callingChat, Function-calling
LicenseApache 2.0Apache 2.0
ReleasedApr 2025Jul 2025

Frequently Asked Questions

Which is better, Qwen3 14B or Qwen3 30B A3B Instruct 2507?

Qwen3 14B and Qwen3 30B A3B Instruct 2507 match in llmrun Score, with values of 97.9 versus 96.5; both fit 24 GB at Q4_K_M; Qwen3 30B A3B Instruct 2507 is 2.5× faster on RTX 4090 (~173 vs ~69 tok/s). Choose Qwen3 14B for fits a 12 GB GPU and lighter on memory; choose Qwen3 30B A3B Instruct 2507 for 2.5× faster on RTX 4090 and longer context window.

Which needs less VRAM, Qwen3 14B or Qwen3 30B A3B Instruct 2507?

At Q4_K_M, Qwen3 14B requires 9.5 GB while Qwen3 30B A3B Instruct 2507 needs 18.7 GB, making Qwen3 14B the easier model to run on your local setup.

Which is faster on RTX 4090?

Estimated decode speed on RTX 4090 is about 69 tok/s for Qwen3 14B (Q4_K_M) and 173 tok/s for Qwen3 30B A3B Instruct 2507 (Q4_K_M), so Qwen3 30B A3B Instruct 2507 is faster. Estimates, ±20%.

Can I run Qwen3 14B or Qwen3 30B A3B Instruct 2507 on a 16 GB GPU?

Qwen3 14B fits at Q4_K_M; Qwen3 30B A3B Instruct 2507 fits at Q3_K_M.

Which has a longer context window, Qwen3 14B or Qwen3 30B A3B Instruct 2507?

Qwen3 14B supports 40,960 tokens and Qwen3 30B A3B Instruct 2507 supports 262,144 tokens.

Qwen3 14B vs Qwen3 30B A3B Instruct 2507 — VRAM & GPU Comparison | llmrun