GLM-5.3 Cost per Task varies ~9x across providers as cache hit rates go live on AA
ArtificialAnlys · x · 2026-10-14
Artificial Analysis launched Cache hit rate and Cost per Task metrics across API providers. For GLM-5.3 (max), Cost per Task varies 9x across 25 providers, driven by the combination of cache hit rate and token pricing: agentic workloads resend the full conversation history every turn, so cacheable input tokens and cache pricing dominate the gap. Fastest output is LithosAI (FP4 ULTRA CHAT) at 714.8 t/s (11x+ spread), lowest latency 3.40s, and cheapest blended price is Supernova (FP4) at $0.41/M tokens (4.4x spread).
Related event: Same Model, 9x Cost Gap Across API Providers, Analysis Shows(2 posts)→
More from Infra
- ASE Panels Spotted at OCP Showfloor, Reportedly Destined for OpenAI's Venice Data Center — BenBajarin · 2026-10-14
- Semicon West and OCP point to next bottlenecks: high-kW liquid cooling, 800VDC, semicap — BenBajarin · 2026-10-14
- Starship launch cadence capped near 145/year at current sites, a bottleneck for orbital compute — JOBhakdi · 2026-10-14
- One Postgres cluster, many previews: Alchemy custom resources give each PR its own database — samgoodwin89 · 2026-10-14
- DeepSeek V4.1 Flash tops Tidelines usage ranking with 5.5T tokens/day, 22.6% share — maferase · 2026-10-14
- Taiwan's AI-boom growth overtakes China's for three straight years — NathanpmYoung · 2026-10-14