AI Cost Calculator
Enter how much you use AI. See what every provider actually costs. API pricing and subscriptions compared side by side.
Pricing checked September 4, 2026 from official sources
API estimates use 500 input and 800 output tokens per message over 30 days. Chat plans are listed separately for reference; their limits and features differ, and they do not include equivalent API usage. Free tiers have usage limits.
How much do you use AI?
Your costs update live below as you drag.
Your monthly cost at 30 messages/day
GPT-6 Luna
CheapestOpenAI
Prompts over 272K tokens are billed at $0.20 / $0.75. Cached input $0.01/M.
GLM-5.3-Flash
Z.ai
Standard rates checked September 9, 2026. The introductory promotion has ended.
GPT-5.6 Luna
OpenAI
GPT-5.4 Nano
OpenAI
DeepSeek V4 Flash
DeepSeek
Peak cache-miss default; off-peak and cache-hit rates are lower.
Qwen3.8-27B
Alibaba Cloud
International Model Studio rate; Singapore lists a 1M-token free quota.
Gemini 3.8 Flash
Intro price through Dec 31, 2026; standard rate is $1.50 / $7.50.
DeepSeek V4 Pro
DeepSeek
Peak cache-miss default; off-peak and cache-hit rates are lower.
GPT-5.4 Mini
OpenAI
Claude Haiku 4.5
Anthropic
Qwen3.8-Max
Alibaba Cloud
International Model Studio rate; region and deployment can change the price.
Grok 4.6
xAI
Short-context rate; long-context requests (>=200K) are $4 / $12.
Gemini 3.5 Flash
Claude Sonnet 5
Anthropic
GPT-6 Sol
OpenAI
Prompts over 272K tokens are billed at $4 / $15. Cached input $0.20/M.
GPT-5.6 Terra
OpenAI
Gemini 3.1 Pro
GPT-5.4
OpenAI
Claude Opus 5.5
Anthropic
Cached input $0.20/M; cache writes $5/M.
GPT-5.6 Sol
OpenAI
$4 / $20 is a promotional rate available at least through Nov 21, 2026; list price is $5 / $30.
Claude Opus 5
Anthropic
Claude Fable 5.1
Anthropic
Cache reads are $0.25/M; plan and cache terms vary.
GPT-6 Astra
OpenAI
API live; organization admins may need to enable it.
GPT-5.5 Pro
OpenAI
Quick take
At 30 messages/day, GPT-6 Luna is the cheapest API option at $0.41/mo. The most expensive option (GPT-5.5 Pro) would cost $143.10/mo for the same usage. That adds up fast.
AI pricing changes constantly
We track price changes and write honest comparisons. Get updates in your inbox.
Full pricing reference (checked September 4, 2026)
| Model | Input/M tokens | Output/M tokens | Subscription |
|---|---|---|---|
Claude Opus 5 | $5 | $25 | Pro ($20/mo), Max (previous generation) |
Claude Opus 5.5 | $4 | $20 | Pro ($20/mo), Max; Claude Code default |
Claude Fable 5.1 | $10 | $50 | API / Max access varies |
Claude Sonnet 5 | $2 | $10 | Pro ($20/mo) |
Claude Haiku 4.5 | $1 | $5 | Pro ($20/mo) |
GPT-6 Astra | $10 | $50 | API available / confirm access |
GPT-6 Sol | $2 | $10 | Plus ($20/mo) and up |
GPT-6 Luna | $0.1 | $0.5 | Free (desktop app) / paid plans |
GPT-5.6 Sol | $4 | $20 | Plus ($20/mo) |
GPT-5.6 Terra | $2 | $12 | API only |
GPT-5.6 Luna | $0.2 | $1.2 | API only |
GPT-5.4 | $2.5 | $15 | Plus ($20/mo) |
GPT-5.5 Pro | $30 | $180 | Pro ($200/mo) |
GPT-5.4 Mini | $0.75 | $4.5 | Plus ($20/mo) |
GPT-5.4 Nano | $0.2 | $1.25 | API only |
Gemini 3.8 Flash | $0.75 | $3.75 | Free tier / paid API |
Qwen3.8-Max | $2 | $6 | API only |
Qwen3.8-27B | $0.5 | $3 | API / open model |
Gemini 3.1 Pro | $2 | $12 | AI Pro ($20/mo) |
Gemini 3.5 Flash | $1.5 | $9 | AI Pro ($20/mo) |
Grok 4.6 | $2 | $6 | Free chat / paid API |
GLM-5.3-Flash | $0.15 | $0.5 | API (standard rate) |
DeepSeek V4 Flash | $0.44 | $1.32 | No subscription |
DeepSeek V4 Pro | $1.32 | $3.96 | No subscription |
Prices from official provider documentation. Anthropic: platform.claude.com. OpenAI: developers.openai.com pricing and model pages. Google: ai.google.dev. Alibaba Cloud: Model Studio pricing. xAI: docs.x.ai. DeepSeek: api-docs.deepseek.com. GPT-6 Sol and Luna rows use the rate for prompts up to 272K tokens. DeepSeek rows use peak cache-miss pricing as a conservative default; cache-hit and off-peak rates are lower. Gemini 3.8 Flash shows Google's introductory rate through December 31, 2026. Regional, long-context, batch, subscription, and workspace prices can differ.