This page may contain stale information. Last updated: 2026-09-23
Definition
AI price wars describe rapid list-price cuts and tier restructuring among frontier LLM API providers competing on cost-per-intelligence for high-volume workloads.
Key Points
-
2026-10-25: Max Woolf timeline—cumulative cuts up to ~80% (Luna Jul 2026), Sep 2026 GPT-6 / Opus 5.5 same-day repricing; cache-read dominates agentic coding spend (2026-10-25-openai-anthropic-llm-price-cuts, 2026-10-26-openai-anthropic-price-cuts-infoworld)
-
2026-09-22: Same-day dual launch — openai gpt-6-sol (10) and gpt-6-luna (0.50) at 50% below GPT-5.6 promo rates; anthropic claude-opus-5-5 at 20 (20% below Opus 5) with Fable 5.1-parity claims (2026-09-22-openai-gpt-6-sol-luna-launch, 2026-09-22-claude-opus-5-5-artificial-analysis)
-
2026-08-16: DeepSeek reverses discount trajectory with peak pricing (up to >4× illustrative Pro workloads) (peak-off-peak-api-pricing)
-
2026-07-30: openai cut gpt-5-6-luna 80% and gpt-5-6-terra 20%; Sol Fast mode at 2× price (2026-07-31-openai-gpt56-price-cuts-official)
-
Competitive pressure cited: Google Gemini Flash-class, Anthropic pricing, Chinese open/cheap models (2026-07-31-openai-gpt56-price-cuts-cnbc)
-
Family spread Sol→Luna expanded to ~25× (ApiDog analysis)
Related
- gpt-6-sol
- gpt-6-luna
- claude-opus-5-5
- anthropic
- openai
- ai-inference
- gpt-5-6-luna
- gpt-5-6-terra
- gpt-5-6-sol
- llm-api-pricing