$17K/month AI API bill: user says flagship model costs are unsustainable
Retromorphix · reddit · 2026-08-21
A developer shares a $17,000 monthly AI API bill, saying tokens-per-task on flagship models have gone through the roof while lower-tier models can't handle most tasks. They're asking the community what actually works to control costs: caching, pruning, compression, bulk buying, or routing to open-source models.
More from Infra
- Agentic CPU Guidebook: Why Agent Workloads Point to a $221B Server CPU Market by 2030 — BenBajarin · 2026-10-16
- Haight Street Compute Market: a one-day trading floor for GPU deals in SF — ctjlewis · 2026-10-16
- Temporal Serverless Workers on Cloud Run: queue-driven autoscaling, scale to zero — rseroter · 2026-10-16
- LMA brings day-0 model support to TPU for the first time, spanning NVIDIA, AMD, Intel via SGLang — ying11231 · 2026-10-16
- Semtech's analyst-day slide cleanly explains the four signal-integrity impairments — BenBajarin · 2026-10-16
- Fortytwo Engine boosts Qwen3.8-27B accuracy 7% and cuts tokens 25% via Swarm Inference — rohanpaul_ai · 2026-10-16