27B

Qwen3.8-27B on the iPhone 18 Neural Engine at 15–70 tok/s (1.5 bpw)

Shapelearn Qwen 3.8 27B (13.1 GB VRAM)

Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint

Benchmarking Qwen 3.8 27B on RTX 5090 and beyond — VRAM capacity alone can't overcome severe software and inference engine bottlenecks

Run Qwen3.8 27B locally: real numbers from my Mac Studio

Qwen3.8-27B can silently corrupt business records

Qwen 3.8 27B Uncensored – Testing the Uncensored Qwen Model

I gave Qwen 3.8 27B a reverse-engineering job and it finished in 30 minutes

Qwen3.8-27B runs frontier-class coding agents and reasoning locally, no cloud API required

Credit report shows Meta keeping $27B off its books through advanced geometry

Google’s Gemma 2 series launches with not one, but two lightweight model options—a 9B and 27B

Google unveils its Gemma 2 series, with a 27B parameter model that can run on a single TPU