In Artificial Analysis speed benchmarks, InclusionAI's Ling 3.0 Flash and DeepSeek's V4.1 Flash placed second and third behind Nemotron 3.5 Super VL