Roboflow
Qwen

Qwen: Qwen3.5 9b

Qwen3.5 9b Overview

Qwen3.5-9B is a 9-billion-parameter multimodal foundation model developed by Alibaba Cloud's Qwen team, released on March 2, 2026 as part of the Qwen3.5 model family. Designed for efficient multimodal reasoning and long-context language tasks, it notably outperforms the older Qwen3-30B, a model more than three times its size, on key benchmarks including GPQA Diamond, IFEval, and LongBench.

The model supports vision-language inputs through an early-fusion multimodal architecture built on a dense hybrid foundation of Gated Delta Networks and Gated Attention. It can also operate in a text-only mode by skipping the vision encoder during inference. It provides a 262,144-token context window (extensible to ~1M tokens via YaRN) and is released under the Apache License 2.0. Within the current AI landscape, Qwen3.5-9B offers a strong balance of capability and efficiency, making it well-suited for multimodal assistants, document analysis, long-context reasoning, and developer-deployed agentic systems.

Qwen3.5 9b Interactive Demo

Results appear here. Add an image or pick an example to run Qwen3.5 9b.

Qwen3.5 9b Details & Performance

Details

Resources

Vision Tasks

CaptioningChart Question AnsweringClassificationDocument Question AnsweringImage TaggingMulti-Label ClassificationOCRObject DetectionVision LanguageVisual Question Answering

Features

Foundation VisionLLMs with Vision CapabilitiesMultimodal Vision

Usage

Past 30 Days

Performance

Avg. Latency

Qwen3.5 9b 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 3, 2026Pricing updated September 3, 2026

Overall score#33 of 52
66.1%
Avg cost / sample#21 of 52
$0.0008
Avg speed / sample#47 of 52
31.33s
Avg tokens / sample
5.7K

Strengths and weaknesses

Qwen3.5 9b averages 66.1% across the six Vision Evals tasks, ranking #33 of 52 models overall.

Its weakest relative showing is OCR, ranking #48 of 52 at 79.1%.

At $0.0008 per sample it is the 21st cheapest of the 52 benchmarked models, and its average inference time of 31.3s per sample makes it the 47th fastest.

Performance profile

Field medianQwen3.5 9b

Field medians: Object Detection 54.1%, Counting 56.8%, Identification 84.4%, OCR 88.1%, Data Extraction 84.5%, Reasoning 53.6%.

Results by task

TaskScoreField (0 to 100)RankCost / sampleSpeed
Object Detection
45.8%
#32 of 52$0.000827.53s
Counting
54.0%
#31 of 52$0.000943.37s
Identification
84.4%
#21 of 52$0.000519.31s
OCR
79.1%
#48 of 52$0.000628.92s
Data Extraction
82.7%
#33 of 52$0.000413.72s
Reasoning
50.3%
#29 of 52$0.001046.90s

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.

51 models on the current benchmark · scores and efficiency pooled across all six tasks at low effort · Qwen3.5 9B highlighted

Qwen3.5 9b scores are from a single run per task; a three-run re-run under the current protocol is pending · Methodology

View all Vision Evals →

Self-hosted benchmarks

3 quantizations of Qwen3.5 9B, 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 at low effort, the same setting as the hosted frontier models. 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.

QuantizationWeightsOverall scoreDetection mAP50Single-stream tok/s
AWQ INT4Ranked
12.4 GB
66.1%
45.8%169.8
19.3 GB
64.8%
44.9%135.5
14.0 GB
64.4%
43.6%169.9

Qwen3.5 9b Pricing

Qwen3.5 9b costs $0.100 per 1M input tokens and $0.150 per 1M output tokens.

Input$0.100 / 1M tokens
Output$0.150 / 1M tokens

Pricing updated Sep 3, 2026

Alternatives to Qwen3.5 9b

Other models worth comparing for similar use cases.

Qwen
Qwen3 VL 8B Instruct
Qwen3 VL 8B Instruct is an open-weight multimodal vision-language model developed by Qwen / Alibaba Cloud as part of the Qwen3-VL series, designed for instruction-following tasks that combine text with visual inputs such as images and video. Released around October 2025 under the Apache-2.0 license, it targets developers who need capable multimodal reasoning without the scale or cost of very large models.The model contains roughly 8.8 billion dense parameters and supports text, image, and video understanding with strong spatial perception, visual reasoning, and emerging visual agent abilities such as GUI interaction. A standout feature is its native ~256K token context window, extendable to around 1M tokens, enabling long-document reading and extended video comprehension. In today’s landscape, it balances openness, long-context capacity, and solid multimodal performance against heavier proprietary models. Typical applications include multimodal assistants, document and video analysis, visual question answering, and research or product prototyping where transparency and deployability matter.
Google
Gemma 3 4B
Gemma 3 4B, released on March 12, 2025, is the mid-sized member of Google DeepMind’s open-weight Gemma 3 family. With about 4 billion parameters, it is multimodal—supporting text and image inputs and generating text outputs. Like the larger Gemma 3 models, it features a 128,000-token input context window with an output capacity of ~8,192 tokens, enabling it to handle long documents and mixed text–image reasoning tasks.The 4B variant is designed as a balance between efficiency and capability: it offers multilingual support across 140+ languages, strong summarization and reasoning performance, and compatibility with moderate hardware. Inference can run with ~6.4 GB VRAM in BF16, or significantly less in quantized 8-bit (~4.4 GB) or 4-bit (~3.4 GB) modes, making it accessible to developers outside large-scale infrastructure. While it lags behind the 12B and 27B versions on the most complex reasoning and multimodal benchmarks, its lower compute footprint makes it ideal for research, prototyping, and practical deployment where efficiency matters.
OpenAI
GPT-5.4 Nano
GPT-5.4 nano is a high-throughput model developed by OpenAI and released on March 17, 2026, as the efficiency-optimized entry in the GPT-5.4 family. Engineered for cost-sensitive production environments and latency-critical workloads, it features an expanded 400,000-token context window that enables the processing of large document batches or extensive logs in a single pass. The model is primarily optimized for text-heavy operations, serving as a premier engine for high-volume classification, data extraction, ranking, and the orchestration of lightweight sub-agents where speed and low per-token costs are the primary requirements.While it supports text and image inputs, GPT-5.4 nano is designed as a text-first worker rather than a specialized visual reasoning tool. In multi-model architectures, it is best utilized for structured text tasks and simple coding sub-tasks, leaving intensive vision reasoning and UI navigation to its sibling, GPT-5.4 mini. Compared to the previous GPT-5 nano, this version provides a significant leap in reliability for structured outputs and tool calling, making it a dependable and economical choice for developers building scalable, automated pipelines that require rapid execution at the edge of the GPT-5.4 ecosystem.
Google
Gemini 3.1 Flash-Lite
Gemini 3.1 Flash-Lite is a natively multimodal reasoning model from Google DeepMind in the Gemini 3 series, based on the Gemini 3 Pro architecture. It processes text, image, video, audio, and PDF inputs within a 1 million token context window and produces text output up to 64K tokens. The model targets high-volume, latency-sensitive workloads and supports visual question answering, image and document data extraction, content moderation, classification, translation, automated speech recognition, and agentic data pipelines. It exposes configurable thinking levels of minimal, low, medium, and high, which set the depth of internal reasoning applied per request and let developers balance response quality against cost and latency.On benchmarks reported at launch, Gemini 3.1 Flash-Lite scores 86.9% on GPQA Diamond and 76.8% on the MMMU Pro multimodal benchmark, and reaches an Elo score of 1432 on the Arena.ai leaderboard. According to Artificial Analysis benchmarks, it produces a 2.5 times faster time to first answer token and a 45% increase in output speed relative to Gemini 2.5 Flash. It also shows improved instruction following, higher audio input quality for automated speech recognition tasks, and support for structured JSON output used in data extraction pipelines.
Anthropic
Claude Haiku 4.5
Claude Haiku 4.5 is Anthropic’s lightweight model in the Claude 4.5 series, released in October 2025 under a proprietary license. Designed for speed and cost efficiency, it delivers near-frontier performance while maintaining Anthropic’s AI Safety Level 2 standard. Haiku 4.5 supports both text and multimodal (text and image) inputs, integrates tool use and extended reasoning, and features a 200,000 token context window, making it adept at handling long or complex workflows. Though the parameter count remains undisclosed, it achieves about 73.3% on SWE-bench Verified, reflecting strong coding and reasoning ability. Haiku 4.5 is ideal for developers and researchers seeking rapid, cost-effective model calls for analysis, coding, or multimodal understanding.
Qwen
Qwen2.5 VL 7B Instruct
Qwen2.5-VL-7B-Instruct is a 7-billion parameter vision-language model from Alibaba’s QwenLM team, released on January 26, 2025 under the Apache 2.0 license. It is the instruction-tuned variant of the 7B scale in the Qwen2.5-VL family, designed to process multimodal inputs such as text, images, charts, documents, and video. The model enables structured outputs—including JSON for structured content and bounding boxes for visual localization. Weights are publicly available on Hugging Face and GitHub, making it suitable for both research and applied multimodal use.

Qwen3.5 9b License

Apache-2.0 · Permissive license

Qwen3.5 9b is released under Apache-2.0, a permissive license. The Qwen3.5 9b 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, Qwen3.5 9b 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 Qwen3.5 9b 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 Qwen3.5 9b?

This is the straightforward case: a permissive license is the best technical solution and you are free to deploy Qwen3.5 9b 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 Qwen3.5 9b Vision

Yes. Qwen3.5 9b 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 Identification at 84.4% (#21 of 52 at low effort). You can test it on your own image in the demo above.

Yes. its transcriptions match the ground truth 79.1% on average (#48 of 52 at low effort) on Vision Evals OCR. Pulling specific fields out of documents (data extraction) scores 82.7%.

Not its strength. On Vision Evals, Qwen3.5 9b scores 45.8% mAP@50 on object detection (#32 of 52 at low effort) and 54.1% judge-graded accuracy on object counting. For production counting or precise localization, pairing it with a specialized detector like RF-DETR or your own trained model in a Roboflow Workflow is usually more reliable: detect the objects, then count the detections.

On our benchmark's task mix, Qwen3.5 9b averages $0.0008 per sample at $0.10 per 1M input and $0.15 per 1M output tokens (#21 of 52 on cost), with an average speed of 31.3s per sample across the benchmark. Actual cost depends on your images and prompts.

On the overall Vision Evals ranking, Qwen3.5 9b sits #33 of 52 at 66.1%, just behind GLM 5.3 Flash (66.3%) and just ahead of Gemini 2.5 Pro (66%). See the full side-by-side: Qwen3.5 9b vs GLM 5.3 Flash.