Weekly deep-research · auto-refreshed
The trackers
Three weekly-refreshed reads on the AI inference stack — the engines, the models, and the silicon underneath. Each version is snapshotted; step through the history down the left of any tracker.
Inference Engines
vLLM, SGLang, TensorRT-LLM, LMDeploy, llama.cpp and friends — versions, trade-offs, throughput.
10 engines · Weekly (Mondays) →
Frontier Models
Frontier closed/open, small & on-device, agentic frameworks, classifiers — benchmarks and pipelines.
43 models & frameworks · Weekly (Wednesdays) →
Inference Silicon
GPUs, wafer-scale, dataflow LPUs, inference ASICs and RISC-V — specs, trade-offs, who is shipping.
12 chips · Weekly (Fridays) →