Ollama v0.32.6 accelerates Qwen3.5 on Apple GPUs, Hugging Face sees trending Qwen3-VL variant

Ollama version 0.32.6 has been released, significantly boosting the performance of the Qwen3.5 model on Apple GPUs through MLX and speculative decoding. This update enhances local AI inference for users with Apple Silicon. Additionally, a new INT8 quantized version of the Qwen3-VL-32B model is trending on Hugging Face, making large multimodal models more accessible for consumer hardware. AI

IMPACT Enhances local AI inference performance on consumer hardware, making powerful models more accessible for users with Apple Silicon.

RANK_REASON This cluster details a software update (Ollama) that improves performance for specific hardware (Apple GPUs) and a trending model variant (Qwen3-VL) with optimization techniques (INT8 quantization), fitting the 'tool' category for software and model accessibility improvements.

Read on dev.to — LLM tag →

• AMD
• Apple GPUs
• ComfyUI
• Hugging Face
• INT8 quantization
• MLX
• NVIDIA
• Ollama
• OpenAI
• Qwen 3.5
• Qwen3-VL-32B

• model release
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →