Alibaba
Qwen3 235B A22b Thinking 2507
The Qwen3-235B-A22B-Thinking-2507 is the latest thinking-enabled model in the Qwen3 series, delivering groundbreaking improvements in reasoning capabilities. This advanced AI demonstrates significantly enhanced performance across logical reasoning, mathematics, scientific analysis, coding tasks, and academic benchmarks—matching or even surpassing human-expert-level performance to achieve state-of-the-art results among open-source thinking models. Beyond its exceptional reasoning skills, the model demonstrates markedly improved general capabilities, including more precise instruction following, sophisticated tool usage, highly natural text generation, and better alignment with human preferences. It also features enhanced 256K long-context understanding, enabling it to maintain coherence and depth across extended documents and complex discussions.
Qwen3 Coder 480B A35B Instructions
Qwen3-Coder-480B-A35B-Instruct is a cutting-edge open-source coding model from Qwen, matching Claude Sonnet’s performance in agentic programming, browser automation, and core development tasks. With native 256K context (extendable to 1M tokens via YaRN), it excels at repository-scale analysis and features specialized function-call support for platforms like Qwen Code and CLINE—making it ideal for complex, real-world development workflows.
Qwen3 235B A22B Instruct 2507
Qwen3-235B-A22B-Instruct-2507 is a multilingual, instruction-tuned mixture-of-experts language model based on the Qwen3-235B architecture, with 22B active parameters per forward pass. It is optimized for general-purpose text generation, including instruction following, logical reasoning, math, code, and tool usage. The model supports a native 262K context length and does not implement "thinking mode" (<think> blocks). Compared to its base variant, this version delivers significant gains in knowledge coverage, long-context reasoning, coding benchmarks, and alignment with open-ended tasks. It is particularly strong on multilingual understanding, math reasoning (e.g., AIME, HMMT), and alignment evaluations like Arena-Hard and WritingBench.
Qwen3 Coder Next FP8
Qwen3-Coder-Next is an open-weight language model specifically designed for coding agents and local development environments. This highly efficient model delivers exceptional performance with only 3 billion active parameters out of a total of 80 billion, achieving results comparable to models with 10 to 20 times more active parameters while maintaining remarkable cost-effectiveness for agent deployment. Through its sophisticated training methodology, Qwen3-Coder-Next excels in advanced agentic capabilities, including long-horizon reasoning, complex tool usage, and robust recovery from execution failures, ensuring reliable performance across dynamic coding tasks. The model’s versatility is further enhanced by its 256k context length and adaptability to various scaffold templates, enabling seamless integration with diverse CLI/IDE platforms such as Claude Code, Qwen Code, Qoder, Kilo, Trae, and Cline, making it an ideal solution for comprehensive development environments.
Qwen3 Next 80B A3B Instruct
Qwen3-Next employs a highly sparse MoE architecture: 80 billion total parameters, but only approximately 3 billion are activated per inference step. Experiments show that, with global load balancing, increasing the total number of expert parameters while keeping the number of activated experts fixed steadily reduces training loss. Compared to Qwen3’s MoE (128 total experts, 8 routed), Qwen3-Next expands to 512 total experts, combining 10 routed experts and 1 shared expert—maximizing resource usage without compromising performance. The Qwen3-Next-80B-A3B-Instruct performs comparably to our flagship model Qwen3-235B-A22B-Instruct-2507, and demonstrates clear advantages in tasks requiring ultra-long context (up to 256K tokens).
Qwen3 Next 80B A3B Thinking
Qwen3-Next employs a highly sparse MoE design: 80 billion total parameters, but only approximately 3 billion are activated per inference step. Experiments show that, with global load balancing, increasing the total number of expert parameters while keeping the number of activated experts fixed steadily reduces training loss. Compared to Qwen3’s MoE (128 total experts, 8 routed), Qwen3-Next expands to 512 total experts, combining 10 routed experts + 1 shared expert — maximizing resource usage without compromising performance. The Qwen3-Next-80B-A3B-Thinking excels at complex reasoning tasks — outperforming higher-cost models like Qwen3-30B-A3B-Thinking-2507 and Qwen3-32B-Thinking, outperforming the closed-source Gemini-2.5-Flash-Thinking on multiple benchmarks, and approaching the performance of our top-tier model Qwen3-235B-A22B-Thinking-2507.
Qwen MT Plus
Qwen-MT is a large language model optimized for machine translation, built on the foundation of the Tongyi Qianwen model. It supports translation across 92 languages—including Chinese, English, Japanese, Korean, French, Spanish, German, Thai, Indonesian, Vietnamese, Arabic, and more—enabling seamless multilingual communication.
Qwen3 32B
It effectively integrates inference and non-inference modes, enabling seamless switching between them during conversations. Its inference performance matches that of QwQ-32B despite having a smaller parameter size, and its overall capabilities significantly exceed those of Qwen2.5-14B, reaching the state-of-the-art (SOTA) level among models of the same scale.
Qwen3 30B A3B
It effectively integrates inference and non-inference modes, enabling seamless switching between them during conversations. Its inference performance matches that of QwQ-32B despite having a smaller parameter size, and its overall capabilities significantly exceed those of Qwen2.5-14B, reaching the state-of-the-art (SOTA) level among models of the same scale.
Qwen3 235B A22B
It effectively integrates inference and non-inference modes, enabling seamless switching between them during conversations. The model’s inference capability significantly outperforms that of QwQ, and its overall performance surpasses that of Qwen2.5-72B-Instruct, reaching the state-of-the-art (SOTA) level among models of the same scale.
Qwen 2.5 7B Instruct
Qwen2.5 is the latest series of Qwen large language models. For Qwen2.5, we are releasing a number of base language models and instruction-tuned language models ranging from 0.5 to 72 billion parameters. Qwen2.5 offers the following improvements over Qwen2: - Significantly more knowledge and greatly improved capabilities in coding and mathematics, thanks to our specialized expert models in these domains. - Significant improvements in following instructions, generating long texts (over 8K tokens), understanding structured data (e.g., tables), and generating structured outputs, particularly JSON. It is more resilient to diverse system prompts, enhancing role-play implementation and condition-setting for chatbots. - Long-context support of up to 128K tokens and the ability to generate up to 8K tokens. - Multilingual support for over 29 languages, including Chinese, English, French, Spanish, Portuguese, German, Italian, Russian, Japanese, Korean, Vietnamese, Thai, Arabic, and more.
Qwen 2.5 VL 72B Instruction Manual
Qwen2.5-VL, the latest vision-language model in the Qwen2.5 series, delivers enhanced multimodal capabilities, including advanced visual comprehension for object and text recognition, chart and layout analysis, and agent-based dynamic tool orchestration. It processes long-form videos (over 1 hour) with key event detection while enabling precise spatial annotation through bounding boxes or coordinate points. The model specializes in extracting structured data from scanned documents (such as invoices and tables) and achieves state-of-the-art performance across multimodal benchmarks covering image understanding, temporal video analysis, and agent task evaluations.
Qwen 2.5 72B Instruct
Qwen 2.5 is the latest series of Qwen large language models. For Qwen 2.5, we are releasing a number of base language models and instruction-tuned language models ranging from 0.5 to 72 billion parameters.
qwen/qwen3-embedding-0.6b
Qwen3 Embedding 8B
The Qwen3 Embedding 8B Model is the latest proprietary embedding model from the Qwen family, specifically optimized for text embedding tasks. Built upon the dense foundational architecture of the Qwen3 series, it fully inherits the base model's exceptional multilingual capabilities, long-context comprehension, and advanced reasoning skills. The Qwen3 Embedding series delivers groundbreaking performance across multiple embedding applications, including text retrieval, code search, text classification, document clustering, and bidirectional text mining.
Wan 2.1: Image to Video
Alibaba Tongyi Wan is renowned for its high image quality, strong temporal consistency, and ability to follow complex prompts, making it ideal for large-scale commercial video generation. Wan 2.1 enhances motion stability and texture detail, making it suitable for bulk production in e-commerce and advertising. Image-to-Video supports driving motion and camera movements using a single reference image, making it ideal for character dance sequences, product demonstrations, and style extensions. The real-time inference API offers stable performance, no waiting time, and affordable pricing.
Wan 2.1 Text to Video
Alitongyi Wan is renowned for its high image quality, strong temporal consistency, and ability to handle complex prompts, making it ideal for large-scale commercial video generation. Wan 2.1 enhances motion stability and texture detail, making it suitable for bulk production in e-commerce and advertising. Text-to-video capabilities allow users to generate storyboards and cinematographic language directly from prompts, enabling rapid prototyping from script to finished video. The real-time inference API offers stable performance, zero wait time, and affordable pricing.
Wan 2.2: Image to Video
Alibaba Tongyi Wan is renowned for its high image quality, strong temporal consistency, and ability to handle complex prompts, making it ideal for large-scale commercial video generation. Wan 2.2 enhances shot continuity and the naturalness of character movements, delivering more stable results in complex scenes. Its image-to-video generation supports driving both motion and camera work using a single reference image, making it suitable for dance performances, product demonstrations, and style extensions. The real-time inference API offers stable performance with no waiting time and is affordably priced.
Wan 2.2 Text to Video
Alibaba Tongyi Wan is renowned for its high image quality, strong temporal consistency, and ability to handle complex prompts, making it ideal for large-scale commercial video generation. Wan 2.2 enhances shot continuity and the naturalness of character movements, delivering more stable results in complex scenes. Text-to-video capabilities allow users to generate storyboards and cinematic language directly from prompts, enabling rapid prototyping from script to finished video. The real-time inference API offers stable performance with no waiting time and is affordably priced.
Wan 2.5 Image-to-Video Preview
Alibaba Tongyi Wan is renowned for its high image quality, strong temporal consistency, and precise prompt adherence, making it ideal for large-scale commercial video generation. Wan 2.5 delivers further improvements in image clarity and prompt adherence, while the preview version facilitates rapid trial-and-error testing. Image-to-video generation supports using a single reference image to drive motion and camera work, making it suitable for dance performances, product demonstrations, and style extensions. The real-time inference API offers stable performance, zero wait times, and affordable pricing.
Wan 2.5 Text-to-Video Preview
Alibaba Tongyi Wan is renowned for its high image quality, strong temporal consistency, and sophisticated prompt adherence, making it ideal for large-scale commercial video generation. Wan 2.5 delivers further improvements in image clarity and prompt adherence, while the preview version facilitates rapid trial-and-error testing. Text-to-video capabilities allow users to generate storyboards and cinematographic styles directly from prompts, enabling quick prototyping from script to finished video. The real-time inference API offers stable performance with no waiting time and is affordably priced.
Z Image Turbo LoRA
The Z Series offers reliable generation capabilities, making it ideal for production environments. Designed for production-level use, this series prioritizes stability and predictable output. It is suitable for general-purpose content generation and tool integration, and can be easily incorporated into your production workflow. The real-time inference API delivers consistent performance with no waiting time and is affordably priced.
Z Image Turbo
The Z Series offers reliable generation capabilities, making it ideal for production environments. Designed for production-level use, this series prioritizes stability and predictable output. It is suitable for general-purpose content generation and tool integration, and can be easily incorporated into your production workflow. The real-time inference API delivers consistent performance with no waiting time and is affordably priced.
Wan 2.6 Video Reference
The Wan2.6 series offers reliable generation capabilities, making it ideal for production environments. Designed for production-grade use, this series prioritizes stability and predictable output. When generating reference videos, it preserves the original video structure while re-rendering the style and texture, making it suitable for fan creations and restoration/upgrades. The real-time inference API delivers stable performance with no waiting time and is affordably priced.
Wan 2.6 Image to Video
The Wan2.6 series offers reliable generation capabilities, making it ideal for production environments. Designed for production-level use, this series prioritizes stability and predictable output. Image-to-video generation supports driving motion and camera work using a single reference image, making it suitable for dance performances, product demonstrations, and stylistic extensions. The real-time inference API delivers stable performance with no waiting time and is affordably priced.
Wan 2.6 Text to Video
The Wan2.6 series offers reliable generation capabilities, making it ideal for production environments. Designed for production-grade use, this series prioritizes stability and predictable output. Its text-to-video capabilities allow users to generate storyboards and cinematographic language directly from prompts, enabling rapid prototyping from script to finished video. The real-time inference API delivers stable performance with no waiting time and is affordably priced.