Phi-3.5-MoE Instruct vs Qwen3-VL-235B-A22B

Phi-3.5-MoE Instruct

Microsoft

Phi-3.5-MoE-instruct is a sparse mixture-of-experts model from Microsoft's Phi research team, released in August 2024 with 42 billion total parameters across 16 experts and approximately 6.6 billion active parameters per forward pass. The model applies Microsoft's small-data, high-quality training philosophy — developed across earlier Phi generations — to a MoE architecture, targeting reasoning quality comparable to much larger dense models at a fraction of the inference compute. Released under the MIT license, it was notable in the research community for demonstrating that MoE efficiency gains could be realized at smaller total parameter counts than typical large-scale MoE deployments.

Alibaba / Qwen

Qwen3-VL-235B-A22B, released by Alibaba's Qwen team in September 2025, is a natively multimodal Mixture-of-Experts large language model with 235 billion total parameters and 22 billion active parameters. It features a 256K token context window (with extrapolation to 1M tokens), native support for text, image, and video input, and joint visual-textual reasoning capabilities. Qwen3-VL-235B targets complex visual reasoning, video understanding, and multimodal agentic tasks under the Apache 2.0 license.