Model Releases
Tulip uses two techniques to lower latency and boost throughput. CUDA graphs pre‑record GPU work so it launches in one call, reducing CPU ov…
Tulip uses two techniques to lower latency and boost throughput. CUDA graphs pre‑record GPU work so it launches in one call, reducing CPU overhead. LazyTensors tracks results asynchronously, letting t
Tulip uses two techniques to lower latency and boost throughput. CUDA graphs pre‑record GPU work so it launches in one call, reducing CPU overhead. LazyTensors tracks results asynchronously, letting the CPU prep the next batch while the GPU finishes the current one.
Related
- We benchmark Qwen3.6-35B-A3B on an M5 Max MacBook Pro. Across ten prompt lengths and ten decode contexts, Lily was faster than MLX-LM, avera…
- We’ve developed our own inference engine Runtime-Optimized Serving Engine (ROSE) to serve models ranging from embeddings to trillion-paramet…
- 28 TPS on Qwen2.5-7B across two separate cloud regions over public WAN using speculative decoding + CUDA Graphs [P]
Source: Perplexity (X) | 2026-09-04