Leaderboard / Alibaba (Qwen)

Qwen3.5 27B

Name: Qwen3.5 27B
Price: 0.688 USD
Author: Alibaba (Qwen)

Alibaba (Qwen) · Qwen3.5 · released 2026-02-24 · data checked 2026-11-01

Qwen3.5 27B is a 27B-parameter dense model from Alibaba (Qwen) (China), released 2026-02-24 under the Apache-2.0 license. It scores 70.6 on the Small LM Index, #7 of 34 medium models. It needs about 17.9 GB of memory at 4-bit with an 8K-token context and costs $0.69 per million tokens through the cheapest API provider.

Small LM Index

70.6

#7 of 34 scored medium models

Likely 69.2–72.0 · from 6 of 6 core benchmarks

Parameters

27B

Dense

Price per 1M tokens

$0.69

$0.25 in / $2.00 out · blended 3:1

Context window

262K

tokens

Benchmarks

Each bar shows this model's score. The tick marks the best medium model on that benchmark. Sources: 8 reported by Alibaba (Qwen). How we label sources.

MMLU-Pro Lab-reported Knowledge86.1

Class best: Qwen3.6 27B 86.2

GPQA Diamond Lab-reported Knowledge85.5

Class best: Nanbeige4.3 33B 89.5

Humanity's Last Exam Lab-reported Reasoning24.3

Class best: Qwen3.8 27B 30.8

AIME Lab-reported Math · AIME 202692.6

Class best: Ternary Bonsai 2 27B 95.8

LiveCodeBench Lab-reported Coding · LiveCodeBench v680.7

Class best: Qwen3.8 27B 90.3

SWE-bench Verified Lab-reported Coding75.0

Class best: Qwen3.6 27B 77.2

Terminal-Bench Lab-reported Agentic · Terminal-Bench 2.041.6

Class best: Nanbeige4.3 33B 80.9

MMMU-Pro Lab-reported Vision75.0

Class best: Gemma 4 31B 76.9

Specifications

Lab: Alibaba (Qwen) (CN · China)
Architecture: Dense transformer
Total parameters: 27B
Active parameters: All (dense)
Size class: Medium
Context window: 262K tokens
Input modalities: text, image, video
Reasoning: Hybrid (thinking on/off)
License: Apache-2.0
Openness: Open weights
Memory to run: ~17.9 GB at 4-bit, 8K context (more sizes)
Hugging Face: 1.8M downloads (30 days) · 1.1K likes
API providers: 4 on OpenRouter
API price: $0.25 in / $2.00 out per 1M (OpenRouter, 2026-11-03)

Weights on Hugging Face ↗Try on OpenRouter ↗

Compare with

Qwen3.5 27B vs Qwen3.8 27B (27B, index 77.0)
Qwen3.5 27B vs Ternary Bonsai 2 27B (27.4B, index 74.6)
Qwen3.5 27B vs Qwen3.6 27B (27B, index 74.0)
Qwen3.5 27B vs Gemma 3 27B (27.4B, index 32.7)
Qwen3.5 27B vs DiffusionGemma 26B-A4B (25.8B, index 52.6)

Ranked in

Sources

https://huggingface.co/Qwen/Qwen3.6-27B

Something wrong here? Report it.

Run it locally

Uses unsloth/Qwen3.5-27B-GGUF (Q4_K_M). Swap the tag for another size: BF16, IQ4_NL, IQ4_XS, Q3_K_M, Q3_K_S, Q4_0, Q4_1, Q4_K_M….

Ollama

ollama run hf.co/unsloth/Qwen3.5-27B-GGUF:Q4_K_M

llama.cpp

llama-server -hf unsloth/Qwen3.5-27B-GGUF:Q4_K_M

Then open http://localhost:8080 for a chat window.

MLX (Mac)

pip install -U mlx-lm
mlx_lm.chat --model mlx-community/Qwen3.5-27B-OptiQ-4bit

Text only; for images use mlx-vlm.

vLLM (GPU server)

vllm serve Qwen/Qwen3.5-27B

Full-precision weights, for a server GPU.

Open in LM Studio

Memory needed: weights, context cache and runtime buffers.

Quantization	4K context	32K context	256K context
16-bit F16	55 GB	57 GB	72 GB
8-bit Q8_0	30 GB	32 GB	47 GB
4-bit Q4_K_M	18 GB	20 GB	35 GB

Check it against your GPU, Mac or phone →