Latest

Every published item, in order · 9305 items
Oct 19
Tomorrow 08:00, the daily lands in your inboxTelegram channel or email. Same screening as this page, no ads.
ModelsBuildersMulti-sourceSelected03:22
LlamaIndex Launches ExtractBench Pro for Long-Tail Document Extraction

The shift to perfect-document scoring changes how extraction systems should be evaluated: one wrong cell fails the whole doc, which mirrors the economics of enterprise workflows where partial accuracy is often unusable. Teams building document pipelines can use the benchmark and its synthetic template pipeline to stress-test models before deployment without exposing private data. For comparison, its hardest cases sit well above what saturated benchmarks like ExtractBench already measure.

Event
Jerry Liu@jerryjliu0Source
Oct 18
ModelsBuildersMulti-sourceSelected09:58
Kimi K3 hits 842 tokens/s decode on NVIDIA GB300 NVL72

The result shows that kernel-level fusion of attention, MoE and communication, rather than more hardware alone, is now the lever for agent-scale inference speed. Teams serving large MoE models can apply the same CuTe DSL decode path in vLLM to cut latency at low batch sizes. It also demonstrates Moonshot's Kimi K3 running efficiently on NVIDIA's newest GB300 rack, a useful datapoint for cross-hardware deployment comparisons.

Event
Official accountvLLM@vllm_projectSource

Only the latest 2000 items are listed; for older ones see the daily archive