SAVRN
Search Contact SAVRN

Open-weight model · Image and text to text

Qwen3.6-27B

by Qwen Qwen/Qwen3.6-27B

Following the February release of the Qwen3.5 series, we're pleased to share the first open-weight variant of Qwen3.6.

Parameters27.8B
Context262,144
Weights55.6 GB
Licenseapache-2.0
AccessOpen weights
Monthly Downloads3.7M

Runs On

What it takes to serve Qwen3.6-27B (27.8B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.

PrecisionWeightsMemory neededCheapest setupPer hourAlso fits
16-bit 55.6 GB 66.7 GB 1x MI300X (192 GB)
Vultr		 $1.85 1x H100 $1.99 · 1x MI325X $2.00
8-bit 27.8 GB 33.3 GB 1x MI300X (192 GB)
Vultr		 $1.85 1x H100 $1.99 · 1x MI325X $2.00
4-bit 13.9 GB 16.7 GB 1x MI300X (192 GB)
Vultr		 $1.85 1x H100 $1.99 · 1x MI325X $2.00

Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.

SAVRN's Notes on Qwen3.6-27B

Two hosts on the SAVRN Index sell this model by the token, so renting and owning can be priced side by side. With 27.8B parameters, images and text in, text out, it needs 66.7 GB at 16-bit, which fits one 192 GB MI300X, the cheapest listed setup at $1.85 an hour on-demand. Eight-bit drops the need to 33.3 GB and 4-bit to 16.7 GB, so one card carries several copies.

Renting: DeepInfra lists $0.32 input and $3.20 output per million tokens, OVHcloud $0.47 and $3.19. Price your monthly token volume both ways before choosing. Apache 2.0 permits commercial use, modification and redistribution, notices kept and changes stated, so a fine-tune is yours to ship. Verify the 262,144-token context, and note the config's architecture is Qwen3_5ForConditionalGeneration, model type qwen3_5, despite the 3.6 name, so confirm your serving stack loads it. No base model or paper is on record.

SAVRN Research, 2026-09-18

Model Card

By Qwen, published under apache-2.0, revision 6a9e13bd6fc8.

[!Note] This repository contains model weights and configuration files for the post-trained model in the Hugging Face Transformers format.

These artifacts are compatible with Hugging Face Transformers, vLLM, SGLang, KTransformers, etc.

Following the February release of the Qwen3.5 series, we're pleased to share the first open-weight variant of Qwen3.6. Built on direct feedback from the community, Qwen3.6 prioritizes stability and real-world utility, offering developers a more intuitive, responsive, and genuinely productive coding experience.

Qwen3.6 Highlights

This release delivers substantial upgrades, particularly in

  • Agentic Coding: the model now handles frontend workflows and repository-level reasoning with greater fluency and precision.
  • Thinking Preservation: we've introduced a new option to retain reasoning context from historical messages, streamlining iterative development and reducing overhead.

For more details, please refer to our blog post Qwen3.6-27B.

Model Overview

Read the full model card (3,023 words)

Configuration

Architecture
Qwen3_5ForConditionalGeneration
Context length (tokens)
262,144
Layers
64
Hidden size
5,120
Feed-forward size
17,408
Attention heads
24
Key/value heads
4
Head dimension
256
Vocabulary size
248,320
Model type
qwen3_5

Identity and Version

Repository
Qwen/Qwen3.6-27B
Publisher
Qwen
Task
Image and text to text
Modality
Image and text
Library
transformers
Parameters
27.8B parameters
Languages
Not stated by the source
Revision
6a9e13bd6fc8f0983b9b99948120bc37f49c13e9
First published
2026-04-21
Last updated
2026-04-24

Files and Weights

29 files, 55.6 GB in total. The weights are 15 files totalling 55.6 GB in safetensors.

Weights15 files · 55.6 GB
Configuration6 files · 117.6 KB
Tokenizer4 files · 22.9 MB
Documentation2 files · 73.9 KB
Other1 file · 7.8 KB
Repository1 file · 1.6 KB
Every file
FileTypeSizeSHA-256
model-00001-of-00015.safetensorsWeights4.0 GB 5f21d4e349ae
model-00002-of-00015.safetensorsWeights3.9 GB 03de44dc7e93
model-00003-of-00015.safetensorsWeights3.9 GB 5c3a68304dab
model-00004-of-00015.safetensorsWeights3.9 GB ba8b0849cb4c
model-00005-of-00015.safetensorsWeights3.9 GB a5abc1d5e958
model-00006-of-00015.safetensorsWeights3.9 GB 160d914e2e47
model-00007-of-00015.safetensorsWeights4.0 GB 0bcd0ce28c7d
model-00008-of-00015.safetensorsWeights3.9 GB 584a0ed8018d
model-00009-of-00015.safetensorsWeights3.9 GB e7e3e1a17a26
model-00010-of-00015.safetensorsWeights3.9 GB e8934789f474
model-00011-of-00015.safetensorsWeights3.9 GB 44e8fe06d2d6
model-00012-of-00015.safetensorsWeights3.9 GB 33c5d7d18e1b
model-00013-of-00015.safetensorsWeights4.0 GB 68db2ebb0323
model-00014-of-00015.safetensorsWeights3.9 GB 26c114fb6d5d
model-00015-of-00015.safetensorsWeights508.7 MB b84b5b1315e8
config.jsonConfiguration4.3 KB —
configuration.jsonConfiguration51 B —
generation_config.jsonConfiguration202 B —
model.safetensors.index.jsonConfiguration112.2 KB —
preprocessor_config.jsonConfiguration390 B —
video_preprocessor_config.jsonConfiguration385 B —
LICENSEDocumentation11.3 KB —
README.mdDocumentation62.6 KB —
chat_template.jinjaOther7.8 KB —
.gitattributesRepository1.6 KB —
merges.txtTokenizer3.4 MB —
tokenizer.jsonTokenizer12.8 MB 5f9e4d4901a9
tokenizer_config.jsonTokenizer16.7 KB —
vocab.jsonTokenizer6.7 MB —

License and Download

License
apache-2.0
Access
Open weights, no gate
Download size
55.6 GB
Download from Qwen

Released by Qwen through ModelScope. Read the license.

Evaluations

Each result is shown as reported, with the conditions its reporter stated. None is a SAVRN measurement. A comparison lines two results up only when their configuration, unit and setup are all stated and identical.

BenchmarkConditionsResultReported byRevisionDate
Idavidrein/gpqa Task diamondMetric diamondComparison conditions not established 87.8 Model Card
Reported by a third party		 Evaluated revision not stated 2026-04-22
MMMU/MMMU_Pro Task mmmu_pro_visionMetric mmmu_pro_visionComparison conditions not established 75.8 Model Card
Reported by a third party		 Evaluated revision not stated 2026-05-15
MathArena/aime_2026 Task MathArena/aime_2026Metric MathArena/aime_2026Comparison conditions not established 94.1 Model Card
Reported by a third party		 Evaluated revision not stated 2026-04-22
MathArena/hmmt_feb_2026 Task MathArena/hmmt_feb_2026Metric MathArena/hmmt_feb_2026Comparison conditions not established 84.3 Model Card
Reported by a third party		 Evaluated revision not stated 2026-04-22
SWE-bench/SWE-bench_Multilingual Task swe_bench_multilingual_%_resolvedMetric swe_bench_multilingual_%_resolvedComparison conditions not established 71.3 Model Card
Reported by a third party		 Evaluated revision not stated 2026-08-10
SWE-bench/SWE-bench_Verified Task swe_bench_%_resolvedMetric swe_bench_%_resolvedComparison conditions not established 77.2 Model Card
Reported by a third party		 Evaluated revision not stated 2026-04-22
ScaleAI/SWE-bench_Pro Task SWE_Bench_ProMetric SWE_Bench_ProComparison conditions not established 53.5 Model Card
Reported by a third party		 Evaluated revision not stated 2026-04-22
TIGER-Lab/MMLU-Pro Task mmlu_proMetric mmlu_proComparison conditions not established 86.2 Model Card
Reported by a third party		 Evaluated revision not stated 2026-04-22
benchflow/skillsbench Task skillsbench_v1_1Metric skillsbench_v1_1Setup SkillsBench Avg5. Evaluated via OpenCode on 78 tasks (self-contained subset, excluding API-dependent tasks); avg of 5 runs. Mapped to the Hub's benchflow/skillsbench task skillsbench_v1_1.Comparison conditions not established 48.2 Qwen3.6-27B model card — Benchmark Results (Language > Coding Agent)
Reported by a third party		 Evaluated revision not stated 2026-04-21
cais/hle Task hleMetric hleComparison conditions not established 24 Model Card
Reported by a third party		 Evaluated revision not stated 2026-04-22
harborframework/terminal-bench-2.0 Task terminalbench_2Metric terminalbench_2Setup Harbor/Terminus-2 harness; 3h timeout, 32 CPU/48 GB RAM; temp=1.0, top_p=0.95, top_k=20, max_tokens=80K, 256K ctx; avg of 5 runs.Comparison conditions not established 59.3 Model Card
Reported by a third party		 Evaluated revision not stated 2026-06-23
internlm/WildClawBench Task avg_costMetric avg_costComparison conditions not established 20.91 WildClawBench
Reported by a third party		 Evaluated revision not stated 2026-08-11
internlm/WildClawBench Task avg_timeMetric avg_timeComparison conditions not established 421 WildClawBench
Reported by a third party		 Evaluated revision not stated 2026-08-11
internlm/WildClawBench Task overallMetric overallComparison conditions not established 43.2 WildClawBench
Reported by a third party		 Evaluated revision not stated 2026-08-11

Memory Requirements

PrecisionWeights in memory
As published55.6 GB
16-bit55.6 GB
8-bit27.8 GB
4-bit13.9 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Hosted Prices

HostInput / outputUnitObserved
DeepInfra$0.32 / $3.20input / output, per million tokensSep 18, 2026
OVHcloud$0.47 / $3.19input / output, per million tokensSep 18, 2026

From the SAVRN Index.

Built on This Model

Compare Qwen3.6-27B

Questions About Qwen3.6-27B

How much GPU memory does Qwen3.6-27B need?

About 66.7 GB at 16-bit and 16.7 GB at 4-bit: the weights (27.8B parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run Qwen3.6-27B on?

At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.

Can I use Qwen3.6-27B commercially?

Yes. Qwen3.6-27B is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.

What is Qwen3.6-27B's context length?

262,144 tokens, from the maximum position embeddings in its published configuration.

Similar Models

Model · Image and text to text

Qwen3.8-27B

Qwen

Following the widespread community adoption of the Qwen3.5 and Qwen3.6 series, we are pleased to introduce Qwen3.8, the most capable generation in the Qwen open-model family to date. Built on the architectural foundation of Qwen3.5, Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks. Qwen3.8-27B brings these advances to a compact, deployment-friendly dense model: a native vision-language model that understands images and videos, with flexible thinking control, designed to carry complex, multi-step tasks through to completion with greater reliability. For streamlined integration, we recommend using Qwen3.8 via APIs. Qwen3.8 can be…

Open weights apache-2.0 27.8B parameters 262,144 tokens transformers
View model

Model · Image and text to text

Qwen3.8-27B-FP8

Qwen

Following the widespread community adoption of the Qwen3.5 and Qwen3.6 series, we are pleased to introduce Qwen3.8, the most capable generation in the Qwen open-model family to date. Built on the architectural foundation of Qwen3.5, Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks. Qwen3.8-27B brings these advances to a compact, deployment-friendly dense model: a native vision-language model that understands images and videos, with flexible thinking control, designed to carry complex, multi-step tasks through to completion with greater reliability. For streamlined integration, we recommend using Qwen3.8 via APIs. Qwen3.8 can be…

Open weights apache-2.0 27.8B parameters 262,144 tokens transformers
View model

Model · Image and text to text

Qwen3.5-27B

Qwen

Over recent months, we have intensified our focus on developing foundation models that deliver exceptional utility and performance. Qwen3.5 represents a significant leap forward, integrating breakthroughs in multimodal learning, architectural efficiency, reinforcement learning scale, and global accessibility to empower developers and enterprises with unprecedented capability and efficiency. For more details, please refer to our blog post Qwen3.5. WMT24++: a harder subset of WMT24 after difficulty labeling and rebalancing; we report the averaged scores on 55 languages using XCOMET-XXL. Empty cells (--) indicate scores not yet available or not applicable. Empty cells (--) indicate scores not…

Open weights apache-2.0 27.8B parameters 262,144 tokens transformers
View model

Model · Image and text to text

Qwen3.8-27B-AWQ-INT4

Cyankiwi

Following the widespread community adoption of the Qwen3.5 and Qwen3.6 series, we are pleased to introduce Qwen3.8, the most capable generation in the Qwen open-model family to date. Built on the architectural foundation of Qwen3.5, Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks. Qwen3.8-27B brings these advances to a compact, deployment-friendly dense model: a native vision-language model that understands images and videos, with flexible thinking control, designed to carry complex, multi-step tasks through to completion with greater reliability. For streamlined integration, we recommend using Qwen3.8 via APIs. Qwen3.8 can be…

Open weights apache-2.0 27.8B parameters 262,144 tokens transformers
View model

Model · Image and text to text

Qwen3.8-27B

Kyle Thomas

Following the widespread community adoption of the Qwen3.5 and Qwen3.6 series, we are pleased to introduce Qwen3.8, the most capable generation in the Qwen open-model family to date. Built on the architectural foundation of Qwen3.5, Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks. Qwen3.8-27B brings these advances to a compact, deployment-friendly dense model: a native vision-language model that understands images and videos, with flexible thinking control, designed to carry complex, multi-step tasks through to completion with greater reliability. For streamlined integration, we recommend using Qwen3.8 via APIs. Qwen3.8 can be…

Open weights apache-2.0 27.8B parameters 262,144 tokens transformers
View model

Model · Image and text to text

Signal-3.8-27B-FP8

Vwdubb

This is Qwen3.8-27B that gets to the answer faster. AgentionAI Signal is a minimally invasive fine-tune of Qwen3.8-27B designed for lower generation latency and better token efficiency. On our held-out general-prompt evaluation, Signal produces 57% fewer answer tokens and uses 52% fewer thinking tokens, while matching or improving the measured answer quality of the base model. The percentages above were measured on the first release. The weights updated on 2026-09-13 trade a little of that reduction for stability; their re-measurement on the same prompt set is in progress and will replace these numbers. The result is substantially faster end-to-end generation: on typical chat prompts…

Open weights apache-2.0 27.8B parameters 262,144 tokens transformers
View model