Qwen3 235B A22B vs Qwen3 235B A22B Instruct 2507

Qwen3 235B A22B and Qwen3 235B A22B Instruct 2507 match in llmrun Score, with values of 99.6 versus 98.5; neither fits 24 GB.

Model A · Alibaba

Qwen3 235B A22B

99.6llmrun Score

235.1B · 22.2B active · 40K context · Apache 2.0 · Apr 2025

ChatFunctions

Model B · Alibaba

Qwen3 235B A22B Instruct 2507

98.5llmrun Score

235.1B · 22.2B active · 256K context · Apache 2.0 · Jul 2025

ChatFunctions

Which should you choose?

Choose Qwen3 235B A22B Instruct 2507 if you want…

Longer context window
256K vs 40K tokens

Qwen3 235B A22B: no clear edge over Qwen3 235B A22B Instruct 2507 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%.

MemoryQwen3 235B A22BQwen3 235B A22B Instruct 2507

8 GB

RTX 4060 · 8 GB tier

Qwen3 235B A22B

Doesn't fit

smallest quant needs 101.8 GB

Qwen3 235B A22B Instruct 2507

Doesn't fit

smallest quant needs 72.4 GB

12 GB

RTX 3060 12GB · 12 GB tier

Qwen3 235B A22B

Doesn't fit

smallest quant needs 101.8 GB

Qwen3 235B A22B Instruct 2507

Doesn't fit

smallest quant needs 72.4 GB

16 GB

RTX 5060 Ti 16GB · 16 GB tier

Qwen3 235B A22B

Doesn't fit

smallest quant needs 101.8 GB

Qwen3 235B A22B Instruct 2507

Doesn't fit

smallest quant needs 72.4 GB

24 GBreference

RTX 4090 · 24 GB tier

Qwen3 235B A22B

Doesn't fit

smallest quant needs 101.8 GB

Qwen3 235B A22B Instruct 2507

Doesn't fit

smallest quant needs 72.4 GB

32 GB

RTX 5090 · 32 GB tier

Qwen3 235B A22B

Doesn't fit

smallest quant needs 101.8 GB

Qwen3 235B A22B Instruct 2507

Doesn't fit

smallest quant needs 72.4 GB

48 GB

RTX A6000 · 48 GB tier

Qwen3 235B A22B

Doesn't fit

smallest quant needs 101.8 GB

Qwen3 235B A22B Instruct 2507

Doesn't fit

smallest quant needs 72.4 GB

Mac 32 GB

Mac Mini M4 · Macs

Qwen3 235B A22B

Doesn't fit

smallest quant needs 101.8 GB

Qwen3 235B A22B Instruct 2507

Doesn't fit

smallest quant needs 72.4 GB

Mac 64 GB

Mac Studio M4 Max · Macs

Qwen3 235B A22B

Doesn't fit

smallest quant needs 101.8 GB

Qwen3 235B A22B Instruct 2507

Doesn't fit

smallest quant needs 72.4 GB

Mac 128 GB

Mac Studio M4 Max · Macs

Qwen3 235B A22B

Q3_K_L · ~21 tok/s

122.4 GB · up to 32K

Qwen3 235B A22B Instruct 2507

Q3_K_L · ~21 tok/s

122.4 GB · up to 32K

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

Qwen3 235B A22B Qwen3 235B A22B Instruct 2507

Overall

Qwen3 235B A22B

99.696–103

Qwen3 235B A22B Instruct 2507

98.595–102

Coding

Qwen3 235B A22B

101.396–107

Qwen3 235B A22B Instruct 2507

101.696–107

Reasoning

Qwen3 235B A22B

99.292–106

Qwen3 235B A22B Instruct 2507

98.193–104

Qwen3 235B A22B · #52 of 100 open models · #33 of 78 that fit ≤192 GB

Qwen3 235B A22B Instruct 2507 · #56 of 100 open models · #37 of 78 that fit ≤192 GB

Benchmarks

6 shared boards: Qwen3 235B A22B leads 2, Qwen3 235B A22B Instruct 2507 leads 3 · Benchmarks imported Oct 2026

BoardQwen3 235B A22BQwen3 235B A22B Instruct 2507

MMLU-Pro

Qwen3 235B A22B

68.2%#41/145 open

IndependentTIGER-Lab

Qwen3 235B A22B Instruct 2507

83.0%#14/145 open

IndependentTIGER-Lab

Aider Polyglot

Qwen3 235B A22B

59.6%#3/18 open

IndependentEpoch AI · Apr 2025

Qwen3 235B A22B Instruct 2507

59.6%#4/18 open

IndependentEpoch AI · Jul 2025

WeirdML

Qwen3 235B A22B

37.3%#29/46 open

IndependentEpoch AI · Apr 2025

Qwen3 235B A22B Instruct 2507

38.7%#26/46 open

IndependentEpoch AI · Jul 2025

DTBench

Qwen3 235B A22B

75.7%#28/68 open

Lab-reportedEpoch AI · Apr 2025

Qwen3 235B A22B Instruct 2507

78.4%#25/68 open

Lab-reportedEpoch AI · Jul 2025

Fiction.LiveBench

Qwen3 235B A22B

67.7%#9/25 open

IndependentEpoch AI · Apr 2025

Qwen3 235B A22B Instruct 2507

52.9%#16/25 open

IndependentEpoch AI · Sep 2025

LMCA

Qwen3 235B A22B

25.0%#27/54 open

Lab-reportedEpoch AI · Apr 2025

Qwen3 235B A22B Instruct 2507

23.5%#30/54 open

Lab-reportedEpoch AI · Jul 2025

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

LMArena Text

Qwen3 235B A22B

1402.0#47/187 open

ArenaLMArena · Oct 2026

Qwen3 235B A22B Instruct 2507

1422.1#32/187 open

ArenaLMArena · Oct 2026

Only Qwen3 235B A22B (4)

Only Qwen3 235B A22B Instruct 2507 (2)

ARC-AGI-21.3%
ARC-AGI11.0%

VRAM by quantization

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

QuantQwen3 235B A22BQwen3 235B A22B Instruct 2507

Q2_K

Qwen3 235B A22B

100 GB85.7 GB file

32K 103 GB · 128K —

Qwen3 235B A22B Instruct 2507

100 GB85.7 GB file

32K 103 GB · 128K 113 GB

Q3_K_S

Qwen3 235B A22B

103 GB101.4 GB file

32K 106 GB · 128K —

Qwen3 235B A22B Instruct 2507

103 GB101.4 GB file

32K 106 GB · 128K 116 GB

Q3_K_M

Qwen3 235B A22B

115 GB112.4 GB file

32K 118 GB · 128K —

Qwen3 235B A22B Instruct 2507

115 GB112.4 GB file

32K 118 GB · 128K 128 GB

Q4_0

Qwen3 235B A22B

not listed

Qwen3 235B A22B Instruct 2507

118 GB133.1 GB file

32K 121 GB · 128K 130 GB

Q4_K_M

Qwen3 235B A22B

142 GB142.2 GB file

32K 145 GB · 128K —

Qwen3 235B A22B Instruct 2507

142 GB142.2 GB file

32K 145 GB · 128K 154 GB

Q5_K_M

Qwen3 235B A22B

168 GB166.8 GB file

32K 171 GB · 128K —

Qwen3 235B A22B Instruct 2507

168 GB166.8 GB file

32K 171 GB · 128K 180 GB

Q6_K

Qwen3 235B A22B

194 GB193.0 GB file

32K 197 GB · 128K —

Qwen3 235B A22B Instruct 2507

194 GB193.0 GB file

32K 197 GB · 128K 207 GB

Q8_0

Qwen3 235B A22B

236 GB249.9 GB file

32K 239 GB · 128K —

Qwen3 235B A22B Instruct 2507

236 GB249.9 GB file

32K 239 GB · 128K 248 GB

15 more quants

IQ2_XS

Qwen3 235B A22B

not listed

Qwen3 235B A22B Instruct 2507

71.0 GB65.6 GB file

32K 74.0 GB · 128K 83.5 GB

IQ2_S

Qwen3 235B A22B

not listed

Qwen3 235B A22B Instruct 2507

74.0 GB66.1 GB file

32K 76.9 GB · 128K 86.4 GB

IQ2_M

Qwen3 235B A22B

not listed

Qwen3 235B A22B Instruct 2507

79.8 GB78.6 GB file

32K 82.8 GB · 128K 92.3 GB

IQ3_XXS

Qwen3 235B A22B

not listed

Qwen3 235B A22B Instruct 2507

91.6 GB93.1 GB file

32K 94.6 GB · 128K 104 GB

IQ3_XS

Qwen3 235B A22B

not listed

Qwen3 235B A22B Instruct 2507

97.5 GB96.9 GB file

32K 100 GB · 128K 110 GB

IQ3_M

Qwen3 235B A22B

not listed

Qwen3 235B A22B Instruct 2507

106 GB107.3 GB file

32K 109 GB · 128K 119 GB

Q3_K_L

Qwen3 235B A22B

121 GB111.2 GB file

32K 124 GB · 128K —

Qwen3 235B A22B Instruct 2507

121 GB111.2 GB file

32K 124 GB · 128K 133 GB

IQ4_XS

Qwen3 235B A22B

127 GB125.5 GB file

32K 130 GB · 128K —

Qwen3 235B A22B Instruct 2507

127 GB125.5 GB file

32K 130 GB · 128K 139 GB

IQ4_NL

Qwen3 235B A22B

not listed

Qwen3 235B A22B Instruct 2507

133 GB133.1 GB file

32K 136 GB · 128K 145 GB

Q4_1

Qwen3 235B A22B

133 GB147.2 GB file

32K 136 GB · 128K —

Qwen3 235B A22B Instruct 2507

133 GB147.2 GB file

32K 136 GB · 128K 145 GB

Q4_K_S

Qwen3 235B A22B

133 GB133.7 GB file

32K 136 GB · 128K —

Qwen3 235B A22B Instruct 2507

133 GB133.7 GB file

32K 136 GB · 128K 145 GB

Q4_K_L

Qwen3 235B A22B

not listed

Qwen3 235B A22B Instruct 2507

144 GB143.1 GB file

32K 147 GB · 128K 157 GB

Q5_0

Qwen3 235B A22B

147 GB161.9 GB file

32K 150 GB · 128K —

Qwen3 235B A22B Instruct 2507

not listed

Q5_K_S

Qwen3 235B A22B

162 GB161.9 GB file

32K 165 GB · 128K —

Qwen3 235B A22B Instruct 2507

162 GB161.9 GB file

32K 165 GB · 128K 175 GB

BF16

Qwen3 235B A22B

471 GB470.3 GB file

32K 474 GB · 128K —

Qwen3 235B A22B Instruct 2507

471 GB470.3 GB file

32K 474 GB · 128K 483 GB

How to run each

Qwen3 235B A22B

Ollama · Q4_K_M

$ollama run hf.co/unsloth/Qwen3-235B-A22B-GGUF:Q4_K_M

All ways to run Qwen3 235B A22B →

Qwen3 235B A22B Instruct 2507

Ollama · Q4_K_M

$ollama run hf.co/unsloth/Qwen3-235B-A22B-Instruct-2507-GGUF:Q4_K_M

All ways to run Qwen3 235B A22B Instruct 2507 →

Specifications

Specifications of Qwen3 235B A22B and Qwen3 235B A22B Instruct 2507
Spec	Qwen3 235B A22B	Qwen3 235B A22B Instruct 2507
Publisher	Alibaba	Alibaba
Parameters	235.1B	235.1B
Active parameters	22B per token	22B per token
Architecture	Qwen3MoeForCausalLM	Qwen3MoeForCausalLM
Context	40K tokens	256K tokens
Modalities	Chat, Function-calling	Chat, Function-calling
License	Apache 2.0	Apache 2.0
Released	Apr 2025	Jul 2025

Frequently Asked Questions

Which is better, Qwen3 235B A22B or Qwen3 235B A22B Instruct 2507?

Qwen3 235B A22B and Qwen3 235B A22B Instruct 2507 match in llmrun Score, with values of 99.6 versus 98.5; neither fits 24 GB. Choose Qwen3 235B A22B Instruct 2507 for longer context window.

Which needs less VRAM, Qwen3 235B A22B or Qwen3 235B A22B Instruct 2507?

At Q4_K_M, Qwen3 235B A22B requires 141.6 GB while Qwen3 235B A22B Instruct 2507 needs 141.6 GB, making Qwen3 235B A22B the easier model to run on your local setup.

Can I run Qwen3 235B A22B or Qwen3 235B A22B Instruct 2507 on a 16 GB GPU?

Qwen3 235B A22B does not fit (smallest quant needs 101.8 GB); Qwen3 235B A22B Instruct 2507 does not fit (smallest quant needs 72.4 GB).

Which has a longer context window, Qwen3 235B A22B or Qwen3 235B A22B Instruct 2507?

Qwen3 235B A22B supports 40,960 tokens and Qwen3 235B A22B Instruct 2507 supports 262,144 tokens.