When you build a chatbot, an AI agent or an automation, you do not pay a monthly fee. You pay per token through the API. The difference between models is huge: the most expensive model can cost hundreds of times more than the cheapest one for the same amount of text.
Premium models
| Model | Input, per 1M tokens | Output, per 1M tokens |
|---|---|---|
| Claude Fable 5.2 | $8 | $40 |
| GPT-6 Astra | $10 | $50 |
| Claude Opus 5.5 | $4 | $20 |
| Gemini 4 Argon | $2 (rises to $4 on Dec 1) | $10 (rises to $20 on Dec 1) |
| Kimi K3 | $3 | $15 |
Mid-range models
| Model | Input, per 1M tokens | Output, per 1M tokens |
|---|---|---|
| Claude Sonnet 5.5 | $2 | $10 |
| GPT-6.1 Sol | $2 | $10 |
| Kimi K3.1 Pro | $2 | $10 |
| Grok 4.7 | $2 | $6 |
| DeepSeek V4 Pro (peak) | $1.32 | $3.96 |
| GPT-6.1 Terra | $1 | $5 |
Fast and cheap models
| Model | Input, per 1M tokens | Output, per 1M tokens |
|---|---|---|
| Gemini 3.8 Flash | $0.75 | $3.75 |
| DeepSeek V4.1 Flash (peak) | $0.30 | $1.20 |
| Claude Haiku 5.5 (up to 100K) | $0.10 | $0.50 |
| GPT-6 Luna | $0.10 | $0.50 |
Claude prices come from Anthropic’s official page, and DeepSeek prices from DeepSeek’s official page. OpenAI, Google, xAI and Moonshot prices come from their price lists as collected by independent price trackers, so confirm them before you plan a budget.
How to read these numbers
- Output tokens are usually 4 to 5 times more expensive than input, so long answers cost more than long questions.
- Most providers give 50% off for batch requests and big discounts for cached input.
- A cheaper model that needs three attempts can cost more than a better model that gets it right the first time.
Which API to choose
For a customer support bot or bulk text processing, start with Claude Haiku 5.5, GPT-6 Luna or DeepSeek V4.1 Flash. For writing, analysis and agents that must be reliable, test Claude Sonnet 5.5, GPT-6.1 Sol and Gemini 4 Argon. Use the premium models only for the hardest steps.
Prices and dates in this article were checked on November 10, 2026. AI companies change them often, so always confirm on the official pricing page before you buy.
Frequently asked questions
What is the cheapest AI API in 2026?
Among current models, Claude Haiku 5.5 and GPT-6 Luna cost $0.10 per million input tokens, and DeepSeek V4.1 Flash costs $0.15 off-peak.
What is a token?
A small piece of text. In English one token is about three quarters of a word. Armenian text usually needs more tokens for the same meaning.
Why will Gemini 4 Argon get more expensive?
Google offered a launch price until November 30, 2026. From December 1, its listed price is set to double.
Read next
- Claude API Pricing 2026: Fable, Opus, Sonnet and Haiku Costs Explained
- AI Subscription Prices Compared, November 2026: ChatGPT, Claude, Gemini, Grok and More
- DeepSeek Review 2026: V4 Pro, V4.1 Flash and the Cheapest AI API Prices
Want to put the new AI models to work in your business? We build AI agents and automations. AI agents and automation
