Alibaba Releases Qwen3.7-Plus AI Model with Multimodal Capabilities at Lower Cost
Alibaba launched Qwen3.7-Plus, a proprietary AI model supporting text, video and imagery at $0.40/$1.60 per million tokens, departing from open-source strategy.
Alibaba released Qwen3.7-Plus, a new large language model that supports text, video and imagery processing at a significantly lower cost than its predecessor. The model is priced at $0.40 per million input tokens and $1.60 per million output tokens, representing a 60% cost reduction compared to the text-only Qwen3.7-Max model released weeks earlier.
The new model marks a strategic shift for Alibaba, as Qwen3.7-Plus is available only through proprietary APIs and commercial licensing rather than as an open-source release. This represents a departure from Alibaba's previous strategy of releasing open-source models in the Qwen family, which had gained adoption among enterprises including U.S. companies like Airbnb.
Qwen3.7-Plus features a 1-million token context window and includes multimodal capabilities that allow it to process visual interfaces, analyze screenshots, and handle video content. The model incorporates a "preserve_thinking" parameter designed to maintain reasoning continuity across multi-step tasks, addressing what developers describe as "state decay" in long-running automated processes.
Benchmark results show the model achieving competitive performance on specialized tasks. On Terminal Bench 2.0, which measures code execution capabilities, Qwen3.7-Plus scored 70.3, outperforming DeepSeek-V4-Pro Max and Google's Gemini-3.1 Pro. On computer vision benchmarks like ScreenSpot Pro, it achieved a score of 79.0.
The pricing positions Qwen3.7-Plus among the more affordable options in the current AI model market, though it remains more expensive than some Chinese competitors like DeepSeek's models. Alibaba has structured the API to be compatible with existing OpenAI frameworks, potentially simplifying integration for enterprises.
The shift to a proprietary model raises compliance considerations for enterprises, as organizations cannot deploy the model locally and must route data through Alibaba Cloud's managed endpoints. This requirement may affect companies with strict data sovereignty or air-gapped infrastructure requirements.