OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?
Web page · https://minimaxir.com/2026/10/cheap-llms
First seen by Oriel on .
AI summary
The article argues that steep LLM price cuts reflect competition from open-weight models and inference providers, improving serving efficiency, and pressure to keep users as models become easier to switch between.
- From June to September 2026, OpenAI and Anthropic cut prices across several model families, with some effective coding-agent costs falling by 70–75% through lower cache-read pricing.
- For coding agents, cached conversation history accounts for 96%–99% of token expenditure and is hit 90%–100% of the time, making cache-read rates a major cost driver.
- DeepSeek Flash V4.1 1017 is hosted by 29 providers on OpenRouter, and several endpoints offer effective input costs within 10% of DeepSeek’s own service.
- The author attributes pricing pressure to interchangeable models, low switching costs, and expanding inference-provider competition, while noting that OpenAI and Anthropic are investing in quality and other moats.
- Lower API prices do not necessarily reduce customer spending: the author reduced a Codex subscription quota after GPT-6.1 Sol made usage roughly four times more efficient; OpenAI also cut usage allowances for its $200/month tier.
#llm pricing #open-weight models #inference providers #llm agents
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Points
Score reported by the source.
- Hacker News: 428, latest observation ; +416 over observed period since
Comments
Comment count reported by the source.
- Hacker News: 430, latest observation ; +423 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hacker News | 50236102 | https://minimaxir.com/2026/10/cheap-llms/ Read discussion |