A practical comparison tool

What will your tokens cost?

Set a workload once, choose the models you care about, and compare what the same usage could cost.

Prices reviewed 19 October 2026

Your scenario

Set the workload

Uses public standard text rates in USD. No API key needed. Your inputs stay in this browser.

The comparison

Monthly estimate

8 models

Choose models to compare
Lowest for this workload—
Highest for this workload—
Difference—
Estimated monthly API cost comparison by selected model
Model / categoryInput / 1MOutput / 1MMonthly total
Claude Haiku 5.5Anthropic · Fast & efficientHigher rate when a request exceeds 100k input tokens. $0.05$0.25 $0.0003Rate source ↗
DeepSeek FlashDeepSeek · Fast & efficientPeak rate; off-peak pricing is lower. $0.30$1.20 $0.0013Rate source ↗
Qwen3.7 PlusAlibaba Cloud · BalancedInternational deployment, standard non-promotional price. $0.40$1.60 $0.0018Rate source ↗
Gemini 3.8 FlashGoogle · Fast & efficientIntroductory price through 31 Dec 2026. $0.75$3.75 $0.0037Rate source ↗
Claude Sonnet 5.5Anthropic · Balanced $1.00$5.00 $0.0050Rate source ↗
DeepSeek V4 ProDeepSeek · AdvancedPeak rate; off-peak pricing is lower. $1.32$3.96 $0.0050Rate source ↗
Gemini 4 ArgonGoogle · AdvancedIntroductory price through 30 Nov 2026; then $4 input / $20 output. $2.00$10.00 $0.0100Rate source ↗
GPT-5.4OpenAI · AdvancedVery long prompts above 272k tokens use a higher rate. $2.50$15.00 $0.0140Rate source ↗

An estimate, not a bill. Standard text input and output tokens only; caching, batch discounts, image/audio/video tokens, taxes, and tool calls are excluded. Tiered rates are applied to the input size of each request. DeepSeek uses peak rates. Check each linked provider rate before budgeting.