Orgs now switch between 12 models a year; blended token cost fell to $0.14
Tiancaixinxin · x · 2026-10-08
New usage data on enterprise AI model adoption:
Model switching accelerates
- Average models used per org rose from 8 in January to 12 in September.
- File read/write, shell execution, DevOps config, frontend UI, debugging, and SQL tasks migrate quickly to new models.
- Security audit, research, finance/trading, math, writing, roleplay, and support users stick to one model.
Market dynamics
- Anthropic (55%) vs OpenAI (45%) competition is fierce.
- Anthropic leads 4-week rolling retention at 24.8% by Week 12.
- Claude Opus 5.5 consolidated Anthropic spend; no outflow window yet.
- GPT Luna and Sol absorb usage from other models; GPT Astra creates entirely new spend.
Cost & open weights
- Agents have higher cache rates; newer models exceed 90% — blended cost per token fell from $0.77 to $0.14 this year.
- Open-weight models grew from 25% to 60% of total usage; 70% of weekly tokens are from Chinese open-weight models.
More from Infra
- Nebius up 160% vs CoreWeave's 10%: the AI infrastructure stock divergence explained — Beth_Kindig · 2026-10-08
- Microsoft's $5,999 Surface RTX Spark Dev Box preorders open, ships November — tomwarren · 2026-10-08
- Together scales open-source inference with IBM and NVIDIA on B300 cluster — togethercompute · 2026-10-08
- How a Self-Appending Summarizer Quadrupled Token Costs in a 1.4M-Conversation Agent — Good_Education4713 · 2026-10-08
- National Compute gifts $100M in compute credits to White House Genesis Mission — typewriters · 2026-10-08
- Ask: Is Qwen 3.8 Flash-Next Worth It Over 3.6 35B-A3B on 64GB RAM + 16GB VRAM? — msalsas · 2026-10-08