Qwen3.7-Plus

The cost tier of Alibaba's Qwen3.7 generation, sitting below the Qwen3.7-Max flagship and succeeding Qwen3.6-Plus. Proprietary and API-only (qwen3.7-plus on Alibaba Cloud Model Studio; also on OpenRouter from June 3), it previewed anonymously on LM Arena as "Qwen3.7-Plus-Preview" from mid-May before general availability on June 1, 2026. Unlike the text-only Max it takes text, image and video input — OpenRouter describes it as "a cost-effective model in Alibaba's Qwen3.7 series... building on the series' text capabilities with a comprehensive upgrade" to vision — with a 1M-token context, up to 131K output tokens, and optional extended thinking (on by default) with a reported 256K-token reasoning budget. List price is $0.40 / $1.60 per million input / output tokens, roughly a sixth of Qwen3.7-Max on input. Parameter count and architecture are undisclosed.

Press reports of Qwen's launch table position it as a GUI/computer-use and long-context specialist: ScreenSpot Pro 79.0 (GPT-5.4 67.4), AndroidWorld 81.0 (Gemini 3.1 Pro 70.7), MRCR-v2 128K 91.7, GPQA Diamond 90.3, and SWE-Bench Pro ~57.6 versus Qwen3.7-Max's 60.6 (grounding evals run with thinking disabled on Qwen's harness). Artificial Analysis scores it 25 on the Intelligence Index v4.3 (Qwen3.7-Max: 29; Qwen3.6-Plus: 27) and notes it is slow for its price tier (~55 tokens/s, ~2.2 s to first token) and verbose (~110M output tokens to run the index versus a 29M median).