Qwen: Qwen3.6 27B

Qwen3.6 27B Overview

Qwen3.6-27B is a dense 27-billion-parameter multimodal language model developed by Alibaba's Qwen team and released on April 22, 2026. It combines a causal language model with an integrated vision encoder, supporting text, image, and video inputs natively. The architecture employs a hybrid attention design that interleaves Gated DeltaNet linear attention blocks with standard Gated Attention layers across 64 transformer layers with a hidden dimension of 5,120. Unlike Mixture-of-Experts variants in the Qwen3.6 family, all 27 billion parameters are active on every inference pass, simplifying deployment and quantization. The model supports a native context window of 262,144 tokens, extensible to approximately 1,010,000 tokens via YaRN scaling. It is released under the Apache 2.0 license with open weights available on Hugging Face and ModelScope.

The model introduces two notable capabilities relative to prior Qwen releases: enhanced agentic coding support covering frontend workflows and repository-level reasoning, and a Thinking Preservation mechanism that retains chain-of-thought reasoning context across multi-turn conversation history to reduce redundant token generation in iterative agent sessions. It supports both a thinking mode for multi-step reasoning and a non-thinking mode for faster responses within a single model. On coding benchmarks, Qwen reports scores of 77.2 on SWE-bench Verified, 59.3 on Terminal-Bench 2.0, and 48.2 on SkillsBench. Vision capabilities include chart understanding (CharXiv RQ: 78.4), OCR (CC-OCR: 81.2), and video understanding (VideoMME with subtitles: 87.7).

Qwen3.6 27B Interactive Demo

Upload an image

Drag and drop an image here, or click to browse

JPEGPNGGIFWebP

Results appear here. Add an image or pick an example to run Qwen3.6 27B.

Qwen3.6 27B Details & Performance

Details

Resources

GitHub

Vision Tasks

CaptioningChart Question AnsweringClassificationDocument Question AnsweringImage TaggingMulti-Label ClassificationOCRVideo ClassificationVision LanguageVisual Question Answering

Features

Foundation VisionLLMs with Vision CapabilitiesMultimodal Vision

Usage

Past 30 Days

Performance

Avg. Latency

Qwen3.6 27B Vision Evals

Vision Evals is Roboflow's ground-truth benchmark: every model runs the same real-world samples across six vision tasks, and answers are scored against ground truth.

Evals updated September 5, 2026Pricing updated September 15, 2026

Overall score#20 of 53

73.6%

Avg cost / sample#27 of 53

$0.0051

Avg speed / sample#52 of 53

42.09s

Avg tokens / sample

4.1K

Strengths and weaknesses

Qwen3.6 27B averages 73.6% across the six Vision Evals tasks, ranking #20 of 53 models overall.

Its weakest relative showing is Identification, ranking #32 of 53 at 82.3%.

At $0.0051 per sample it is the 27th cheapest of the 53 benchmarked models, and its average inference time of 42.1s per sample makes it the 52nd fastest.

Performance profile

Field medianQwen3.6 27B

Field medians: Object Detection 53.9%, Counting 56.8%, Identification 84.4%, OCR 88.7%, Data Extraction 84.5%, Reasoning 54.1%.

Results by task

Task	Score	Rank	Cost / sample	Speed
Object Detection	59.7% ±0.9, Mean of 3 runs, range 59.0 to 60.8	#15 of 53	$0.0023	17.47s
Counting	67.1% ±4.7, Mean of 3 runs, range 62.2 to 71.6	#18 of 53	$0.0092	78.17s
Identification	82.3% ±1.6, Mean of 3 runs, range 81.3 to 84.4	#32 of 53	$0.0052	44.57s
OCR	88.5% ±1.9, Mean of 3 runs, range 86.7 to 90.6	#28 of 53	$0.0044	36.22s
Data Extraction	84.5% ±1.0, Mean of 3 runs, range 83.5 to 85.6	#25 of 53	$0.0026	19.21s
Reasoning	59.2% ±1.7, Mean of 3 runs, range 57.6 to 60.9	#21 of 53	$0.0095	80.78s

Price vs. performance

Score vs. cost

Overall benchmark score against estimated cost per sample, on a log scale. Upper-left is the sweet spot: high quality at low cost.

52 models on the current benchmark · scores and efficiency pooled across all six tasks at low effort · Qwen3.6 27B highlighted

Qwen3.6 27B scores are the mean of 3 runs per task at low effort; this model has no separate high-effort tier because its chat template only switches thinking on or off · Methodology

View all Vision Evals →

Self-hosted benchmarks

3 quantizations of Qwen3.6 27B, served with vLLM from the published weights and scored on the same tasks as the hosted models. Expand a row for the GPUs it was measured on.

Self-hosted rows run with thinking on, the same setting as the hosted frontier models. Rows marked with a run count are the mean of three runs per task under the benchmark protocol; the rest are single runs awaiting their re-run. Some hosted open-weight rows still run without thinking, so a self-hosted quant can score above its own hosted API. Quantization still costs a little precision, and scores vary between runs.

Quantization	Weights	Overall score	Detection mAP50	Single-stream tok/s
BF16	55.6 GB	71.5%	58.0%	46.5
AWQ INT4	20.4 GB	71.5%	61.1%	95.2
FP83 runsRanked	30.9 GB	73.6%	59.7%	71.6

Qwen3.6 27B Pricing

Qwen3.6 27B costs $0.300 per 1M input tokens and $2.00 per 1M output tokens.

Input$0.300 / 1M tokens

Output$2.00 / 1M tokens

Cached input$0.030 / 1M tokens

Pricing updated Sep 15, 2026

Alternatives to Qwen3.6 27B

Other models worth comparing for similar use cases.

Qwen3.8 27B

Qwen3.8-27B is a dense vision-language model of roughly 27.78 billion parameters from Alibaba's Qwen team, built on the architectural foundation established by the Qwen3.5 generation. It takes interleaved text, image, and video input through an integrated vision encoder and produces text, covering document and chart understanding, visual question answering, front-end code generation from visual references, and agentic computer-use workflows. The decoder is a hybrid stack of 64 layers that repeats a pattern of three Gated DeltaNet linear-attention blocks, each followed by a feed-forward block, then one grouped-query full-attention block, so only a quarter of the layers maintain a conventional key-value cache. Native context length is 262,144 tokens.Post-training targets long-horizon agentic work, and the model exposes per-request thinking control that lets callers switch reasoning traces on or off and choose a reasoning effort level. Multi-token prediction weights are trained alongside the main network for speculative decoding. Qwen reports improvements over Qwen3.6-27B on agentic and multimodal evaluations including Terminal-Bench 2.1, OSWorld-Verified, and SWE-bench Multimodal, with additional results on MathVision, CharXiv, and an internal Vision2Web suite.

Qwen3.6 35B A3B

Qwen3.6-35B-A3B is a sparse Mixture-of-Experts (MoE) multimodal language model developed by the Qwen team at Alibaba Group. It carries 35 billion total parameters but activates only approximately 3 billion per forward pass via a learned routing mechanism, giving it the representational capacity of a large dense model at a fraction of the inference compute. The model is natively multimodal, processing images, documents, and video alongside text as a core architectural capability rather than an add-on. It supports a native context window of 262,144 tokens, extensible up to 1,010,000 tokens via YaRN. A key design feature is the unified thinking/non-thinking mode framework: users can switch between deliberate chain-of-thought reasoning and fast direct responses within a single model, and a "thinking preservation" option retains reasoning context across multi-turn agentic workflows to reduce redundant computation.The model is specifically optimized for agentic coding tasks, including repository-level reasoning, frontend workflow generation, multi-step tool use, and MCP (Model Context Protocol) integration. On SWE-bench Verified it scores 73.4%, on Terminal-Bench 2.0 it scores 51.5%, and on MCPMark it scores 37.0%. For vision-language tasks it achieves 92.0 on RefCOCO, 89.9 on OmniDocBench 1.5, and 83.7 on VideoMMMU. The model also supports Multi-Token Prediction (MTP) for speculative decoding. All Qwen3.6 open-weight models are released under the Apache 2.0 license.

Gemma 4 31B

Gemma 4 31B is the largest dense model in Google's Gemma 4 family, built from the same research as Gemini 3 and released as open weights under the Apache 2.0 license. It supports a 256K token context window with text and image input, configurable thinking mode for step-by-step reasoning, and multilingual support across 140+ languages. The unquantized model fits on a single 80GB GPU.For vision tasks, Gemma 4 31B supports image understanding with variable aspect ratios and resolutions, and can output structured bounding boxes for UI element detection, making it useful for document parsing and UI understanding. Compared to Gemma 3, it delivers stronger reasoning and multimodal performance. It is part of a four-size family alongside the 26B A4B MoE variant and two on-device models (E2B, E4B), with the 31B dense variant optimized for output quality and fine-tuning over inference speed.

Gemma 4 26B A4B

Gemma 4 26B A4B is the Mixture-of-Experts variant in Google's Gemma 4 family, with 25.2B total parameters but only 3.8B active per token. Built from the same Gemini 3 research as the 31B dense sibling and released as open weights under the Apache 2.0 license, it supports a 256K token context window with text and image input and configurable thinking mode. The "A4B" in the name refers to its approximately 4B active parameters. The MoE design makes it significantly faster at inference than the dense 31B, running nearly as fast as a 4B-parameter model while delivering roughly 97% of the dense model's quality.For vision tasks, the 26B A4B shares the same multimodal capabilities as the 31B image understanding with variable aspect ratios and resolutions, and structured bounding box output for UI element detection. The tradeoff versus the 31B dense model is a small quality reduction in exchange for much faster inference and lower hardware requirements, fitting in 18GB of VRAM at 4-bit quantization. It ranked #6 among open models on the Arena AI text leaderboard at launch.

Mistral Small 3.1 24B

Mistral Small 3.1 24B, released on March 17, 2025, is an open-weight multimodal model from Mistral AI, distributed under the Apache-2.0 license. With around 24B parameters and a 128K token context window, it is available in both base and instruction-tuned (“Instruct”) variants. The model introduces vision support alongside text, enabling tasks like multimodal reasoning, captioning, and image-based Q&A.It is multilingual, supporting many languages, and is optimized for fast responses, function calling, structured dialogue, and long-context reasoning. Despite its size, the model can be run locally in quantized formats, fitting on machines with ~32GB RAM, making it accessible to developers outside large cloud setups. However, the output length is smaller than the 128K input window, meaning long generations may require chaining. In addition, using full vision features or the maximum context window significantly increases compute costs, and performance on highly complex reasoning or enterprise-scale tasks still trails larger proprietary frontier models.

Qwen3.6 27B License

Apache-2.0 · Permissive license

Qwen3.6 27B is released under Apache-2.0, a permissive license. The Qwen3.6 27B license lets you run, fine-tune, and redistribute the model in commercial products with no obligation to open-source related code changes, so no separate commercial license is required.

Commercial use: Permitted. Because Apache-2.0 is permissive, Qwen3.6 27B can ship inside paid products and internal systems with no commercial license and no revenue threshold.
Modification: Permitted. Fine-tuning, quantizing, and distilling are all allowed, and your code changes can stay closed. Files you change must be marked as changed.
Redistribution: Permitted with attribution. Ship the Apache-2.0 license text and any NOTICE file alongside the weights or derived code.

Apache-2.0 grants an express patent license that terminates if you bring a patent claim over the work, and it disclaims warranties. Validate Qwen3.6 27B on your own data before you depend on it in production.

Read the full Apache 2.0 license ↗

Do I need a commercial license for Qwen3.6 27B?

This is the straightforward case: a permissive license is the best technical solution and you are free to deploy Qwen3.6 27B commercially without open-sourcing your own code.

Do not hesitate to reach out with questions for your commercial project — our team will help you start solving business problems on the first call. See Roboflow commercial licensing for the models included in each plan.

Talk to sales

This model is released under the Apache License 2.0, a permissive open-source license that allows commercial use, modification, distribution, and patent use.

Yes. Under the terms of the Apache 2.0 license, you can freely use this model for commercial purposes, including in proprietary products. You must retain the copyright notice and disclaimers when redistributing.

License information is provided as a guide and is not legal advice.

Frequently Asked Questions About Qwen3.6 27B Vision

Yes. Qwen3.6 27B accepts image input and handles OCR, data extraction, object counting, identification, visual reasoning, and object detection. On Roboflow's Vision Evals its strongest task is Object Detection at 59.7% (#15 of 53 at low effort). You can test it on your own image in the demo above.

Yes. its transcriptions match the ground truth 88.5% on average (#28 of 53 at low effort) on Vision Evals OCR. Pulling specific fields out of documents (data extraction) scores 84.5%.

It's serviceable. On Vision Evals, Qwen3.6 27B scores 59.7% mAP@50 on object detection (#15 of 53 at low effort) and 67.1% judge-graded accuracy on object counting.

On our benchmark's task mix, Qwen3.6 27B averages $0.0051 per sample at $0.30 per 1M input and $2.00 per 1M output tokens (#27 of 53 on cost), with an average speed of 42.1s per sample across the benchmark. Actual cost depends on your images and prompts.

On the overall Vision Evals ranking, Qwen3.6 27B sits #20 of 53 at 73.6%, just behind GPT-5.6 Terra (73.8%) and just ahead of Qwen3.6 35B-A3B (71.9%). See the full side-by-side: Qwen3.6 27B vs GPT-5.6 Terra.