LLM API Pricing Calculator
Enter your monthly token volume, pick the models you're weighing up, and see the monthly bill for each — side by side, from published rate cards.
1 · Your monthly token volume
2 · Models to compare
3 · Monthly cost, side by side
| Model | Input rate /1M | Output rate /1M | Input cost | Output cost | Monthly total | vs cheapest |
|---|
Not sure how many tokens your text is? Use the token counter to measure real prompts first, or check the VRAM calculator if you're weighing self-hosting an open-weight model instead.
The full rate table (July 2026)
Every rate below links to the model page it was taken from. Prices are USD per 1 million tokens, input / output.
| Model | Input /1M | Output /1M | Context window |
|---|---|---|---|
| DeepSeek V4-Flash | $0.14 | $0.28 | 1M tokens |
| Kimi K2.6 | $0.60 | $2.50 | 262,144 tokens |
| Claude Haiku 4.5 | $1.00 | $5.00 | 200,000 tokens |
| Gemini 3.5 Flash | $1.50 | $9.00 | 1,048,576 tokens |
| DeepSeek V4-Pro | $1.74 | $3.48 | 1M tokens |
| Claude Sonnet 5 (intro, to 31 Aug 2026) | $2.00 | $10.00 | 1M tokens |
| Gemini 3.1 Pro (approx.) | ~$2.00 | ~$12.00 | 2M tokens (3.1 family) |
| Grok 4.5 | $2.00 | $6.00 | 500,000 tokens |
| Qwen 3.7-Max | $2.50 | $7.50 | 1M tokens |
| Claude Sonnet 5 (standard, from 1 Sep 2026) | $3.00 | $15.00 | 1M tokens |
| Claude Opus 4.8 | $5.00 | $25.00 | 1M tokens |
| Claude Opus 4.7 | $5.00 | $25.00 | 1M tokens |
| GPT-5.5 (Thinking) | $5.00 | $30.00 | 1M tokens (400K in Codex CLI) |
| Claude Fable 5 | $10.00 | $50.00 | 1M tokens |
| Claude Opus 4.8 (fast mode) | $10.00 | $50.00 | 1M tokens |
| GPT-5.5 Pro | $30.00 | $180.00 | 1M tokens |
Frequently asked questions
How is LLM API cost calculated?
Monthly cost = (input tokens ÷ 1,000,000 × input price per 1M) + (output tokens ÷ 1,000,000 × output price per 1M). This calculator uses published list prices only; prompt-caching, batch and cache-hit discounts can lower real bills substantially (e.g. Claude Opus 4.7's page lists up to 90% savings with prompt caching).
Which LLM API is cheapest in 2026?
On the list prices tracked here, DeepSeek V4-Flash ($0.14/$0.28 per 1M) is the lowest, followed by Kimi K2.6 ($0.60/$2.50) and Claude Haiku 4.5 ($1/$5). Cheapest per token is not always cheapest per solved task — quality and retry rates matter.
Why are some models missing?
We never estimate prices. Models without a publicly confirmed USD rate card are excluded: GPT-5.6 (preview, pricing TBD), GPT-Live (voice models in ChatGPT — API pricing TBD), Mistral Small 4 (pricing TBD), Grok 5 (unreleased), Doubao Seed 2.1 (priced in CNY — ¥6/¥30 per 1M — with no verified USD card), Gemma 4 (open weights, no first-party API price) and Cursor 3 (an IDE with subscription tiers, not a per-token API).
Do these prices include caching or batch discounts?
No — list price only. Several vendors publish separate discounted paths (Gemini 3.5 Flash cached input at $0.15/1M; Doubao cache-hit input from ¥1.2/1M; Anthropic prompt caching and batch discounts). If your workload is cache-friendly, treat these numbers as an upper bound.