Nemotron compared to 3 other models on deep context performance

View on Reddit

Benchmark on Strix Halo 128GB shows Nemotron's prompt processing degrades less in deep context vs gpt-oss 120B and Qwen. Uses Lemonade Server and Vulkan backend. Full comparison data in gallery.