ServiceNow-AI/SuperApriel-15B-Instruct · Hugging Face
ServiceNow released SuperApriel-15B-Instruct, a single-checkpoint 15B model with 8 deployment presets spanning 1.0× to 10.7× decode throughput at 32K sequence length. It has 48 decoder layers with 4 mixer variants per layer and up to 262K context positions depending on runtime; the key point is that speed-quality tradeoffs and speculative decoding are exposed from the same weights.
Why it matters: ServiceNow released SuperApriel-15B-Instruct, a 15B instruction model whose single weight set offers eight speed tiers, up to 10x faster decode throughput, so latency and quality trade off inside one model.