Alibaba Qwen · Released August 2026
Qwen3.8 27B
What it takes to run Qwen3.8 27B locally: memory by quantization, the GPUs that handle it, and how fast it runs.
Updated
- Parameters
- 27B
- Architecture
- Dense
- Context
- 256K tokens
- License
- Apache 2.0 license
- Memory at Q4_K_M
- ~18GB
What do you need to run Qwen3.8 27B?
Qwen3.8 27B needs about 18GB of VRAM at Q4_K_M with 8K context (16GB for the weights plus KV cache and overhead). The smallest GPU that can hold it is a 16GB card like the RTX 4060 Ti 16GB at Q3_K_M, with little memory to spare. For a good balance of price and speed we recommend the RTX 3090 (~39 tokens/sec).
What to buy
GPUs for Qwen3.8 27B
Speeds are estimates at the quantization shown with 8K context. We may earn a commission when you buy through these links. Learn more.
Memory
Qwen3.8 27B VRAM requirements by quantization
Lower-bit quantizations shrink the model at some cost to quality. Q4_K_M is the usual sweet spot. Totals include the KV cache for the context length shown plus runtime overhead.
| Quantization | Quality | Weights | Total @ 8K ctx | Total @ 32K ctx |
|---|---|---|---|---|
| Q2_K | Rough | 10 GB | 12 GB | 13 GB |
| Q3_K_M | Usable | 13 GB | 15 GB | 16 GB |
| Q4_K_M | RecommendedDefault | 16 GB | 18 GB | 20 GB |
| Q5_K_M | Very good | 20 GB | 22 GB | 23 GB |
| Q6_K | Near-lossless | 23 GB | 25 GB | 26 GB |
| Q8_0 | Practically lossless | 30 GB | 32 GB | 33 GB |
| BF16 | Full precision | 56 GB | 58 GB | 60 GB |
KV cache note: Hybrid layout: 48 of 64 layers use Gated DeltaNet linear attention with a fixed-size state, and only the 16 full-attention layers keep a growing cache: 2 x 16 x 4 KV heads x 256 dims x 2 bytes = 65,536 bytes (64 KiB) per token, about 17GB at the full 262K context.
Performance
How fast Qwen3.8 27B runs
Single-user generation speed on popular GPUs that fit the model entirely in memory, at the best quantization each can hold.
RTX 5090
Q6_K · 32GB
~59tok/s
RTX 5080
Q3_K_M · 16GB
~55tok/s
RTX 5070 Ti
Q3_K_M · 16GB
~51tok/s
RTX 4090
Q4_K_M · 24GB
~47tok/s
RTX PRO 6000
Q8_0 · 96GB
~46tok/s
RTX 4080 Super
Q3_K_M · 16GB
~43tok/s
RTX 3090
Q4_K_M · 24GB
~39tok/s
RTX 4070 Ti Super
Q3_K_M · 16GB
~39tok/s
RX 7900 XTX
Q4_K_M · 24GB
~36tok/s
RTX PRO 5000 72GB
Q8_0 · 72GB
~35tok/s
Measured
Qwen3.8 27B benchmarks
- 4× RTX 3090Q5_K_M · llama.cpp39 tok/s llama.cpp discussion #15013: Performance of llama.cpp on NVIDIA CUDA
- RX 9060 XT 16GBIQ4_XS · llama.cpp20 tok/s llama.cpp discussion #10879: Performance of llama.cpp with Vulkan
- RX 9060 XT 16GBIQ4_XS · llama.cpp20 tok/s llama.cpp discussion #15021: Performance of llama.cpp on AMD ROCm (HIP)
Overview
About Qwen3.8 27B
Qwen3.8 27B, released in August 2026, is one of the strongest open models that fits a single 24GB GPU, with image input, a 262K native context, and switchable thinking. It uses the same hybrid layout as the Qwen3.5 and 3.6 27B models, where three of every four layers use linear attention, so long contexts cost far less memory than on older dense models. At Q4_K_M it needs about 18GB with 8K context, leaving room for 64K or more of context on a 24GB card, and 32GB cards can run Q6_K. It supersedes Qwen3.6 27B and Qwen3 32B, but its long reasoning traces make speed on slower hardware a real consideration.
Strengths
- +Top-tier reasoning and agentic coding for a model that fits a 24GB GPU
- +Image input, toggleable thinking, and 262K native context
- +Hybrid attention keeps the KV cache at 64 KiB per token
- +Apache 2.0 license
Weaknesses
- −Uses many output tokens when thinking, so answers take a while
- −Dense model: slower than 3B-active mixture-of-experts models on the same GPU
- −Does not fit 16GB GPUs at Q4_K_M without offloading
Setup
How to run Qwen3.8 27B
$ ollama run qwen3.8:27b$ llama-server -hf unsloth/Qwen3.8-27B-GGUF:Q4_K_M -c 8192 -ngl 99Weights: Qwen/Qwen3.8-27B · GGUF: unsloth/Qwen3.8-27B-GGUF
Alternatives
Models like Qwen3.8 27B
Alibaba Qwen
Qwen3.6 35B-A3B
35B (3B active) · 256K ctx
Gemma 4 31B
31B · 256K ctx
Alibaba Qwen
Qwen3 32B
32.8B · 32K ctx
Alibaba Qwen
Qwen3.5 122B-A10B
122B (10B active) · 256K ctx
Frequently asked questions
How much VRAM do you need to run Qwen3.8 27B?
About 18GB at Q4_K_M with 8K tokens of context: 16.5GB of weights, 0.5GB of KV cache, and 1.1GB of runtime overhead. Higher-precision quants and longer contexts need more; see the table above.
What is the cheapest GPU that can run Qwen3.8 27B?
The smallest card in our database that can hold it is the RTX 4060 Ti 16GB (16GB), at Q3_K_M with little memory to spare, generating ~17 tokens/sec.
Is Qwen3.8 27B free for commercial use?
Qwen3.8 27B is released under the Apache 2.0 license. That is a permissive license, so commercial use is allowed.
How long a context does Qwen3.8 27B support?
Up to 262,144 tokens (256K). Long contexts need a lot of extra memory for the KV cache, so most local setups run 8K to 32K.
Sources: Unsloth Qwen3.8 27B GGUF files, modelfit.io GGUF file sizes for Qwen3.8 27B, Artificial Analysis: Qwen3.8 27B vs Qwen3.6 35B A3B. Capability score basis: Artificial Analysis Intelligence Index: 52 before the v4.3 rescale (August 2026) and 33.7 on v4.3 (September 2026), placed below the open frontier models that score 42 to 44 on v4.3.