Canada vs. ... Alabama?

A detailed comparison of four open-source models—Qwen 3.8 27B, Nemotron Lightning 30B, Ornith 35B, and Muse-Glimmer 30B—was conducted using real-world coding tests. Qwen 3.8 27B emerged as the clear winner in coding and architecture, producing 50,000 tokens in 25 minutes and performing flawlessly in computer use MCP tests, comparable to GPT-5.6 Luna. Ornith demonstrated impressive capabilities and speed, while Nemotron and Muse-Glimmer lagged behind in the overall performance evaluations conducted via llm-bench.io.