Cost Per Token Over Time: The AI Price Collapse
How AI cost per token fell roughly 10x per year since 2021, why it collapsed, and why your per-task bill did not drop nearly as fast.
The price of intelligence
AI costs
Token prices are the sticker. Context, retries, subscriptions and unfinished work decide the bill.
- In this world
- 80 analyses
- Field tool
- Model your cost per task ↗
Topic archive · page 3
Continue the descent.
Analysis of AI pricing, API bills, subscriptions and the real cost of completing useful work.
How to Use the Kimi K3 API: Access, Code, Limits
Call Kimi K3 through Moonshot OpenAI-compatible endpoint or OpenRouter. Get a key, set the base URL, use model kimi-k3, and mind the max reasoning setting.
Kimi K3 vs DeepSeek V4: Price, Specs, Benchmarks
Kimi K3 and DeepSeek V4 are both open-weight Chinese models. K3 leads the AA intelligence index; DeepSeek V4 costs a fraction and leads on SWE-bench.
Who Owns Moonshot AI? Backers, Valuation, Money
Moonshot AI, the Chinese lab behind Kimi, is founder-led by Yang Zhilin, with Alibaba its largest backer and a reported 20 billion dollar valuation.
Inkling: Thinking Machines' First Open-Weight Model
Thinking Machines released Inkling, a 975B open-weight model under Apache 2.0. Full pricing, benchmarks against GLM-5.2 and DeepSeek V4, and the verdict.
OpenRouter AI Explained: Pricing, Privacy and Tradeoffs
OpenRouter AI gives developers one API for 400+ models. See its real fees, privacy controls, pros, cons, and when going direct costs less.
What Does CoreWeave Do? AI Cloud Business Explained
What does CoreWeave do? See how its GPU cloud works, who uses it, how it makes money, its NVIDIA relationship, and the risks behind its rapid growth.
AI Data Center Financing 2026: Inside the $700B Buildout
How AI infrastructure gets financed in 2026: Helix (KKR/Nvidia), Apollo/Blackstone (Anthropic), Stargate (OpenAI), and what this means for model pricing.
Best Open-Weight AI Models in 2026
The best open-weight AI models in 2026, ranked by use case: coding, long context, multimodal, on-device, and the real cost per finished task.