Alibaba releases Qwen3.5-Medium models rivaling Claude Sonnet 4.5 on consumer GPUs

Alibaba's Qwen team released the Qwen3.5 Medium Model series: four new LLMs including three open-source models (Qwen3.5-35B-A3B, Qwen3.5-122B-A10B, Qwen3.5-27B) under Apache 2.0. The flagship Qwen3.5-35B-A3B exceeds 1M token context length on consumer GPUs with 32GB VRAM. Benchmarks show the open-source models beating OpenAI's GPT-5-mini and Anthropic's Claude Sonnet 4.5, while maintaining accuracy when quantized. A fourth proprietary model, Qwen3.5-Flash, offers steep cost advantages over Western competitors via Alibaba Cloud's API.

View full digest for February 26, 2026