Inference providers on OpenRouter sharply improved agentic cache rates in 10 months
AccBalanced · x · 2026-11-05
A chart by PeterJWalker shows inference providers across the OpenRouter ecosystem making major gains in agentic cache hit rate over roughly 10 months. Each bubble is one provider: x-axis is cache rate, y-axis is tokens served, bubble size is number of models served.
More from Infra
- SGLang Summit Lands in SF Nov 12-13: 2,000+ Builders, Production Lessons From NVIDIA, AMD, Alibaba and More — ying11231 · 2026-11-05
- SGLang lands twice in GitHub Octoverse 2026 Top 10 as agents drive record 17.9B pushes — ying11231 · 2026-11-05
- LoopVQ Hits Sub-1-bit KV Cache Quantization With Up to 6.1x Throughput — SeoulNationalUniv · 2026-11-05
- Hark on scaling multimodal data curation with SGLang and KubeRay, plus faster inference-worker startup — tianyin_xu · 2026-11-05
- Open-Source hunch Distills LLM Judgments Onto CPU: Millisecond Decisions Under $0.03/Million Tokens — aigclink · 2026-11-05
- SpaceX targets first orbital AI data centers for late 2027, scale in 2028 — mitchdeg · 2026-11-05