Top open LLM downloads on HF drop ~30%; hypothesis: developers shifting to OpenRouter gateways
reach_vb · x · 2026-10-11
Nathan Lambert observed a decline in Hugging Face downloads across top open LLMs: many leading models saw a 30% drop in download rate that is continuing over time, possibly partly due to a change in HF's counting.
@reachvb offers a hypothesis: developers/tinkerers are increasingly using bigger open-weight models via OpenRouter and other gateways — hosting 500B+ parameter models is non-trivial, and labs' cheaper, faster small models have eroded most of the utility of running a small model locally.
The takeaway: downloads no longer equal actual usage; inference gateways are siphoning the open-source model entry point.
Related event: Top open LLM downloads drop ~30% on Hugging Face, suspected metric change(7 posts)→
More from Infra
- Speech Model Shrunk 13x to 153M Params by Looping 2 Shared Blocks — pbaylies · 2026-10-11
- Inference demand went vertical, yet is a flat line next to post-training/RL growth — zainhas · 2026-10-11
- Pat Gelsinger slams HBM as "a lousy memory" wasting four bits for every one it makes — SumitGup · 2026-10-11
- Qualcomm CEO predicts AI phone supercycle, smart glasses as top AI wearable — SuB8u · 2026-10-11
- Hugging Face launches a PyTorch profiling series: from torch.profiler to attention — ariG23498 · 2026-10-11
- Bain sees 183GW of new data center capacity by 2030, needing $5-6.5T in spending — Beth_Kindig · 2026-10-11