Qwen3.5 Series
Enhanced Qwen3.5 with native multimodal, 1M context window, and agentic capabilities.
Q
Qwen
Alibaba's multilingual LLM with excellent Chinese-English performance
📊 Rating Breakdown
✅ Pros
- Outstanding Chinese language support
- Wide range of model sizes for different needs
- Completely open source
- Strong coding capabilities especially in Python
- Apache license for commercial use
- Built-in tool use and function calling
- Multilingual support for many languages
- Integrates with Alibaba Cloud services
❌ Cons
- Documentation can be inconsistent
- English performance slightly behind top models
- Less community resources than Llama
- Less brand recognition globally
- Limited multimodal capabilities
- Setup requires more technical knowledge
- Customer support could be better
- Smaller plugin ecosystem
🧠 Available Models
Qwen3.5-Plus
Open SourceAlibaba's latest open-source flagship (2026). Top global open-source performance, matching closed-source leaders in coding and reasoning.
Parameters: Not disclosed
Context: 128K tokens
📝 text 🖼️ image
✅ Top open-source ✅ Coding ✅ Multilingual ✅ Cost-effective
Input: ¥0.80/1M tokens (~$0.11)
Output: ¥2.00/1M tokens (~$0.28)
🔌 API Availablevia Alibaba Cloud API, DashScope
Qwen3-Max
ProprietaryAlibaba's largest model with 1T+ parameters. Outperforms GPT-5 and Claude Opus 4 on LMArena. Released Sep 2025 at Apsara Conference.
Parameters: 1T+ (MoE)
Context: 128K tokens
📝 text 🖼️ image 🎤 audio 🎬 video
✅ Massive scale ✅ Multimodal ✅ LMArena top-3 ✅ Full modality
Input: ¥4.00/1M tokens (~$0.55)
Output: ¥12.00/1M tokens (~$1.65)
🔌 API Availablevia Alibaba Cloud API, DashScope
Qwen3-Plus
Open SourceMid-tier model offering excellent balance of performance and cost. Great for most production workloads.
Parameters: ~70B
Context: 128K tokens
📝 text
✅ Great value ✅ Strong Chinese ✅ Production ready
Input: ¥0.80/1M tokens (~$0.11)
Output: ¥2.00/1M tokens (~$0.28)
🔌 API Availablevia Alibaba Cloud API, DashScope
Qwen-VL-Max
ProprietaryAlibaba's vision-language model for image and video understanding. Supports document OCR, chart analysis, and video QA.
Parameters: Not disclosed
Context: 128K tokens
📝 text 🖼️ image 🎬 video
✅ Document analysis ✅ Video understanding ✅ OCR ✅ Chart reading
Input: ¥3.00/1M tokens (~$0.41)
Output: ¥9.00/1M tokens (~$1.24)
🔌 API Availablevia Alibaba Cloud API, DashScope
Qwen2.5-Coder
Open SourceState-of-the-art open-source code model. Strong across 92 programming languages. Available in 0.5B to 32B sizes.
Parameters: 0.5B / 1.5B / 3B / 7B / 14B / 32B
Context: 128K tokens
📝 text 💻 code
✅ Code generation ✅ Multi-size ✅ Open source ✅ Multiple languages
Input: Free / Open Source
Output: Free / Open Source
🔌 API Availablevia Alibaba Cloud API, Hugging Face, Ollama
API Quickstart Guide
Visit the official API documentation to get started with this tool.
🧠 Available Models
Qwen3.5-Plus
Open SourceAlibaba's latest open-source flagship (2026). Top global open-source performance, matching closed-source leaders in coding and reasoning.
Parameters: Not disclosed
Context: 128K tokens
📝 text 🖼️ image
✅ Top open-source ✅ Coding ✅ Multilingual ✅ Cost-effective
Input: ¥0.80/1M tokens (~$0.11)
Output: ¥2.00/1M tokens (~$0.28)
🔌 API Availablevia Alibaba Cloud API, DashScope
Qwen3-Max
ProprietaryAlibaba's largest model with 1T+ parameters. Outperforms GPT-5 and Claude Opus 4 on LMArena. Released Sep 2025 at Apsara Conference.
Parameters: 1T+ (MoE)
Context: 128K tokens
📝 text 🖼️ image 🎤 audio 🎬 video
✅ Massive scale ✅ Multimodal ✅ LMArena top-3 ✅ Full modality
Input: ¥4.00/1M tokens (~$0.55)
Output: ¥12.00/1M tokens (~$1.65)
🔌 API Availablevia Alibaba Cloud API, DashScope
Qwen3-Plus
Open SourceMid-tier model offering excellent balance of performance and cost. Great for most production workloads.
Parameters: ~70B
Context: 128K tokens
📝 text
✅ Great value ✅ Strong Chinese ✅ Production ready
Input: ¥0.80/1M tokens (~$0.11)
Output: ¥2.00/1M tokens (~$0.28)
🔌 API Availablevia Alibaba Cloud API, DashScope
Qwen-VL-Max
ProprietaryAlibaba's vision-language model for image and video understanding. Supports document OCR, chart analysis, and video QA.
Parameters: Not disclosed
Context: 128K tokens
📝 text 🖼️ image 🎬 video
✅ Document analysis ✅ Video understanding ✅ OCR ✅ Chart reading
Input: ¥3.00/1M tokens (~$0.41)
Output: ¥9.00/1M tokens (~$1.24)
🔌 API Availablevia Alibaba Cloud API, DashScope
Qwen2.5-Coder
Open SourceState-of-the-art open-source code model. Strong across 92 programming languages. Available in 0.5B to 32B sizes.
Parameters: 0.5B / 1.5B / 3B / 7B / 14B / 32B
Context: 128K tokens
📝 text 💻 code
✅ Code generation ✅ Multi-size ✅ Open source ✅ Multiple languages
Input: Free / Open Source
Output: Free / Open Source
🔌 API Availablevia Alibaba Cloud API, Hugging Face, Ollama
API Quickstart Guide
Visit the official API documentation to get started with this tool.
📜 Version History
Sep 2025
Qwen3.5
Sep 2024
Qwen2.5
Qwen2.5 Series
Released Qwen2.5 with 0.5B to 72B sizes, significantly improved coding and mathematical abilities.
Sep 2023
Qwen-14B
Qwen-14B Release
Released Qwen-14B with improved reasoning, knowledge, and multilingual performance.
Mar 2025
Qwen3-110B
Qwen3-110B Launch
Released Qwen3-110B with Mixture-of-Experts architecture, rivaling top-tier models.
Jun 2024
Qwen2-72B
Qwen2 Launch
Released Qwen2 series with 72B flagship, achieving state-of-the-art results on multilingual benchmarks.
Feb 2026
Qwen4
Qwen4 Launch
Next-gen Qwen4 with breakthrough reasoning, multilingual understanding, and enterprise platform.
Feb 2024
Qwen1.5
Qwen1.5 Series
Released Qwen1.5 with multi-model sizes from 0.5B to 72B, improved performance across the board.
Dec 2024
QwQ-32B
QwQ-32B Reasoning Model
Released QwQ-32B reasoning model with advanced chain-of-thought capabilities.
Dec 2023
Qwen-72B
Qwen-72B Launch
Launched Qwen-72B flagship model, competing with GPT-3.5 on multiple benchmarks.
Aug 2023
Qwen-7B
Qwen-7B Launch
Alibaba Cloud launched Qwen-7B, an open-source large language model with strong Chinese and English capabilities.