Qwen3.5 35B A3B vs Qwen3.8 27B
Qwen3.8 27B leads llmrun Score 121.5 to 105.6; both fit 24 GB at Q4_K_M; Qwen3.5 35B A3B is 4.0× faster on RTX 4090 (~151 vs ~38 tok/s).
Model A · Alibaba
Qwen3.5 35B A3B36.0B · 4.7B active · 256K context · Apache 2.0 · Feb 2026
Model B · Alibaba
Qwen3.8 27B27.8B · 256K context · Apache 2.0 · Aug 2026
Choose Qwen3.5 35B A3B if you want…
4.0× faster on RTX 4090
~151 vs ~38 tok/s
Choose Qwen3.8 27B if you want…
Higher llmrun Score
121.5 vs 105.6
Stronger at reasoning
124.1 vs 105.6
Ahead on the benchmarks
6 of 6 shared boards
Fits a 12 GB GPU
IQ3_XXS · 12 GB; Qwen3.5 35B A3B needs 15.7 GB at its smallest quant
Lighter on memory
17.4 vs 22.0 GB at Q4_K_M
Also does function-calling
Qwen3.5 35B A3B doesn't list it
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
Qwen3.5 35B A3B
Doesn't fit
smallest quant needs 15.7 GB
Qwen3.8 27B
Doesn't fit
smallest quant needs 8.8 GB
12 GB
RTX 3060 12GB · 12 GB tier
Qwen3.5 35B A3B
Doesn't fit
smallest quant needs 15.7 GB
Qwen3.8 27B
IQ3_XXS · ~20 tok/s
12 GB · up to 16Kfits
16 GB
RTX 5060 Ti 16GB · 16 GB tier
Qwen3.5 35B A3B
Q2_K · ~108 tok/s
15.7 GB · up to 32K
Qwen3.8 27B
Q3_K_L · ~19 tok/s
15.4 GB · up to 16Kbetter quant
24 GBreference
RTX 4090 · 24 GB tier
Qwen3.5 35B A3B
Q4_K_M · ~151 tok/s
22 GB · up to 128Kfaster
Qwen3.8 27B
Q4_K_M · ~38 tok/s
17.9 GB · up to 64K
32 GB
RTX 5090 · 32 GB tier
Qwen3.5 35B A3B
Q4_K_M · ~213 tok/s
22 GB · up to 256Kfaster
Qwen3.8 27B
Q4_K_M · ~67 tok/s
17.9 GB · up to 256K
48 GB
RTX A6000 · 48 GB tier
Qwen3.5 35B A3B
Q4_K_M · ~125 tok/s
22 GB · up to 256Kfaster
Qwen3.8 27B
Q4_K_M · ~29 tok/s
17.9 GB · up to 256K
Mac 32 GB
Mac Mini M4 · Macs
Qwen3.5 35B A3B
Q4_K_M · ~24 tok/s
22 GB · up to 256Kfaster
Qwen3.8 27B
Q4_K_M · ~5 tok/s
17.9 GB · up to 128K
Mac 64 GB
Mac Studio M4 Max · Macs
Qwen3.5 35B A3B
Q4_K_M · ~64 tok/s
22 GB · up to 256Kfaster
Qwen3.8 27B
Q4_K_M · ~22 tok/s
17.9 GB · up to 256K
Mac 128 GB
Mac Studio M4 Max · Macs
Qwen3.5 35B A3B
Q4_K_M · ~64 tok/s
22 GB · up to 256Kfaster
Qwen3.8 27B
Q4_K_M · ~22 tok/s
17.9 GB · up to 256K
Reference GPU: RTX 4090. Tinted cell = the better pick on that machine (fits, more headroom, or at least 10% faster).
Overall
Qwen3.5 35B A3B
Qwen3.8 27B
Coding
Qwen3.5 35B A3B
no data
Qwen3.8 27B
Math
Qwen3.5 35B A3B
no data
Qwen3.8 27B
Science
Qwen3.5 35B A3B
Qwen3.8 27B
Reasoning
Qwen3.5 35B A3B
Qwen3.8 27B
Qwen3.5 35B A3B · #36 of 92 open models · #5 of 34 that fit 24 GB
Qwen3.8 27B · #10 of 92 open models · #1 of 34 that fit 24 GB
Benchmarks
6 shared boards: Qwen3.5 35B A3B leads 0, Qwen3.8 27B leads 6 · Benchmarks imported Oct 2026
Qwen3.5 35B A3B
1393.8#53/186 open
ArenaLMArena · Oct 2026
Qwen3.8 27B
1438.1#22/186 open
ArenaLMArena · Oct 2026
Qwen3.5 35B A3B
1253.9#46/50 open
ArenaLMArena · Oct 2026
Qwen3.8 27B
1592.8#10/50 open
ArenaLMArena · Oct 2026
Qwen3.5 35B A3B
29.3%#52/65 open
Lab-reportedEpoch AI · Feb 2026
Qwen3.8 27B
46.6%#19/65 open
Lab-reportedEpoch AI · Aug 2026
Qwen3.5 35B A3B
0.6%#38/65 open
Lab-reportedEpoch AI · Feb 2026
Qwen3.8 27B
5.4%#15/65 open
Lab-reportedEpoch AI · Aug 2026
Qwen3.5 35B A3B
80.0%#23/68 open
Lab-reportedEpoch AI · Feb 2026
Qwen3.8 27B
88.0%#9/68 open
Lab-reportedEpoch AI · Aug 2026
Qwen3.5 35B A3B
29.5%#22/54 open
Lab-reportedEpoch AI · Feb 2026
Qwen3.8 27B
41.4%#7/54 open
Lab-reportedEpoch AI · Aug 2026
12 more boards (other shared and single-model boards)
Only Qwen3.5 35B A3B (4)
- GPQA Diamond83.5%
- AIME 2024/202570.0%
- MMLU-Pro85.3%
- Chess Puzzles10.0%
Only Qwen3.8 27B (8)
- LiveBench Coding75.7
- LiveBench Reasoning80.0
- LiveBench Math86.2
- ARC-AGI-242.4%
- ARC-AGI87.5%
- APEX-Agents47.5%
- Surface Evolver Bench45.0%
- ProofBench16.0%
VRAM by quantization
Memory to load the model and 2K context; below it, the total at 32K and 128K context. Lighter side in bold.
Q2_K
Qwen3.5 35B A3B
15.7 GB~15.3 GB file
32K 15.9 GB · 128K 16.9 GB
Qwen3.8 27B
12.6 GB10.8 GB file
32K 13.9 GB · 128K 19.3 GB
Q3_K_S
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
12.9 GB12.7 GB file
32K 14.2 GB · 128K 19.6 GB
Q3_K_M
Qwen3.5 35B A3B
17.9 GB~17.5 GB file
32K 18.2 GB · 128K 19.2 GB
Qwen3.8 27B
14.3 GB13.4 GB file
32K 15.6 GB · 128K 21.0 GB
Q4_0
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
14.6 GB~13.9 GB file
32K 16.0 GB · 128K 21.4 GB
Q4_K_M
Qwen3.5 35B A3B
22.0 GB~21.6 GB file
32K 22.2 GB · 128K 23.2 GB
Qwen3.8 27B
17.4 GB16.5 GB file
32K 18.8 GB · 128K 24.1 GB
Q5_K_M
Qwen3.5 35B A3B
26.0 GB~25.6 GB file
32K 26.3 GB · 128K 27.3 GB
Qwen3.8 27B
20.5 GB19.8 GB file
32K 21.9 GB · 128K 27.3 GB
Q6_K
Qwen3.5 35B A3B
30.0 GB~29.7 GB file
32K 30.3 GB · 128K 31.3 GB
Qwen3.8 27B
23.7 GB22.0 GB file
32K 25.0 GB · 128K 30.4 GB
Q8_0
Qwen3.5 35B A3B
36.3 GB~36.0 GB file
32K 36.6 GB · 128K 37.6 GB
Qwen3.8 27B
28.5 GB29.0 GB file
32K 29.9 GB · 128K 35.2 GB
17 more quants
IQ2_XXS
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
8.4 GB7.3 GB file
32K 9.7 GB · 128K 15.1 GB
IQ2_XS
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
9.1 GB9.1 GB file
32K 10.4 GB · 128K 15.8 GB
IQ2_S
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
9.4 GB8.4 GB file
32K 10.8 GB · 128K 16.1 GB
IQ2_M
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
10.1 GB10.5 GB file
32K 11.5 GB · 128K 16.8 GB
IQ3_XXS
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
11.5 GB10.9 GB file
32K 12.9 GB · 128K 18.2 GB
IQ3_XS
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
12.2 GB10.4 GB file
32K 13.6 GB · 128K 18.9 GB
IQ3_S
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
12.6 GB12.0 GB file
32K 13.9 GB · 128K 19.3 GB
IQ3_M
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
13.3 GB14.9 GB file
32K 14.6 GB · 128K 20.0 GB
Q3_K_L
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
15.0 GB14.1 GB file
32K 16.3 GB · 128K 21.7 GB
IQ4_XS
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
15.7 GB14.3 GB file
32K 17.0 GB · 128K 22.4 GB
IQ4_NL
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
16.4 GB17.4 GB file
32K 17.7 GB · 128K 23.1 GB
Q4_1
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
16.4 GB17.5 GB file
32K 17.7 GB · 128K 23.1 GB
Q4_K_S
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
16.4 GB15.4 GB file
32K 17.7 GB · 128K 23.1 GB
Q4_K_L
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
17.8 GB18.8 GB file
32K 19.1 GB · 128K 24.5 GB
Q5_K_S
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
19.9 GB18.7 GB file
32K 21.2 GB · 128K 26.6 GB
Q5_K_L
Qwen3.5 35B A3B
not listed
Qwen3.8 27B
20.9 GB21.5 GB file
32K 22.2 GB · 128K 27.6 GB
BF16
Qwen3.5 35B A3B
72.3 GB~71.9 GB file
32K 72.5 GB · 128K 73.6 GB
Qwen3.8 27B
56.3 GB54.7 GB file
32K 57.7 GB · 128K 63.0 GB
Qwen3.5 35B A3B
No local recipe · Q4_K_M
No GGUF repo entry exists for this model yet.
All ways to run Qwen3.5 35B A3B →Qwen3.8 27B
Ollama · Q4_K_M
ollama run hf.co/unsloth/Qwen3.8-27B-GGUF:Q4_K_M| Spec | Qwen3.5 35B A3B | Qwen3.8 27B |
|---|---|---|
| Publisher | Alibaba | Alibaba |
| Parameters | 36.0B | 27.8B |
| Active parameters | 4.7B per token | — |
| Architecture | Qwen3_5MoeForConditionalGeneration | Qwen3_5ForConditionalGeneration |
| Context | 256K tokens | 256K tokens |
| Modalities | Vision | Vision, Function-calling |
| License | Apache 2.0 | Apache 2.0 |
| Released | Feb 2026 | Aug 2026 |
Frequently Asked Questions
Which is better, Qwen3.5 35B A3B or Qwen3.8 27B?
Qwen3.8 27B leads llmrun Score 121.5 to 105.6; both fit 24 GB at Q4_K_M; Qwen3.5 35B A3B is 4.0× faster on RTX 4090 (~151 vs ~38 tok/s). Choose Qwen3.5 35B A3B for 4.0× faster on RTX 4090; choose Qwen3.8 27B for higher llmrun Score and stronger at reasoning.
Which needs less VRAM, Qwen3.5 35B A3B or Qwen3.8 27B?
At Q4_K_M, Qwen3.5 35B A3B requires 22.0 GB while Qwen3.8 27B needs 17.4 GB, making Qwen3.8 27B the easier model to run on your local setup.
Which is faster on RTX 4090?
Estimated decode speed on RTX 4090 is about 151 tok/s for Qwen3.5 35B A3B (Q4_K_M) and 38 tok/s for Qwen3.8 27B (Q4_K_M), so Qwen3.5 35B A3B is faster. Estimates, ±20%.
Can I run Qwen3.5 35B A3B or Qwen3.8 27B on a 16 GB GPU?
Qwen3.5 35B A3B fits at Q2_K; Qwen3.8 27B fits at Q3_K_L.
Which has a longer context window, Qwen3.5 35B A3B or Qwen3.8 27B?
Qwen3.5 35B A3B supports 262,144 tokens and Qwen3.8 27B supports 262,144 tokens.