How is the cost worked out?

Cost per call is input tokens times the model's input price, plus output tokens times its output price, divided by a million. The month is calls a day times 30. It uses each vendor's list price, with no discounts, no prompt caching and no batch pricing.

Why is Jev so much cheaper?

Jev bills input tokens only, at $0.042 per million, and charges nothing for output, because it returns a typed answer (yes or no, a choice, a score) instead of text. A chat model bills for every token it writes. When your task is a decision, that difference is most of the bill.

Can Jev replace the other models here?

Only for decisions. It cannot write, summarize or answer in prose. Tick the box that says your calls need written text, and Jev leaves the comparison. For the jobs it does suit, see how it compares in the published runs on the Jev vs an LLM page.

Are the prices current?

They were copied from each vendor's pricing page on 2026-10-02, and each model name links to its source. Vendors change prices, and Google lists a rise for Gemini 3.8 Flash from 1 January 2027, so check the source before you commit a budget.

Why do the same tokens cost differently by model?

Tokenizers differ. The same text can be a different number of tokens on different models, and Anthropic says its newer tokenizer produces about 30% more tokens for the same text. The calculator takes your token counts at face value, so measure them on each model you compare.

Does this tool call a model or save anything?

No. It is arithmetic in your browser. The numbers live only in the page link, so you can share them.

LLM Cost Calculator: Compare API Costs Per Month

What will your LLM calls cost?

Enter your calls a day and tokens per call. You get the monthly cost on 10 models, from vendor list prices. Jev is in the list: it bills input only, so a decision costs a fraction of what a chat model charges.

Calls per day

Input tokens per call

Output tokens per call

My calls need written text back (an answer, a summary, code). Jev returns typed answers, not text, so it drops out of the comparison.

Jev costs $6.30 a month at these numbers. The cheapest other model here costs $16.50, the dearest $660.

Monthly and per-1,000-call cost by model, cheapest first
Model	Per 1,000 calls	Per month
Jev TypeSafe $0.042 in / $0 out per million Returns typed answers (yes/no, a choice, a score), not text. Output is free.	$0.02	$6.30
GPT-6 Luna OpenAI $0.1 in / $0.5 out per million	$0.06	$16.50
GPT-4o mini OpenAI $0.15 in / $0.6 out per million	$0.08	$24.30
Gemini 3.5 Flash-Lite Google $0.3 in / $2.5 out per million	$0.17	$52.50
Gemini 3.8 Flash Google $0.75 in / $3.75 out per million Google lists this price through 31 December 2026, then $1.50 input and $7.50 output from 1 January 2027.	$0.41	$124
GPT-5.4 mini OpenAI $0.75 in / $4.5 out per million	$0.42	$126
Claude Haiku 4.5 Anthropic $1 in / $5 out per million	$0.55	$165
GPT-6.1 Sol OpenAI $2 in / $10 out per million	$1.10	$330
Claude Sonnet 5.5 Anthropic $2 in / $10 out per million	$1.10	$330
Claude Opus 5.5 Anthropic $4 in / $20 out per million	$2.20	$660

30 days a month. List prices, no discounts, no caching.

How to estimate LLM API cost per month

Multiply three numbers: calls a day, tokens per call, and the model’s price per million tokens, once for input and once for output. A chat model bills both. A classifier or a moderation check sends a few hundred tokens in and gets a few back, so input dominates. A chatbot reply is the reverse: output costs several times more per token and there is more of it.

The prices this LLM cost calculator uses

US dollars per million tokens, standard tier, copied from each vendor on 2026-10-02. Click a model for its source.

Model	Input	Output
Jev TypeSafe	$0.042	$0
GPT-6 Luna OpenAI	$0.1	$0.5
GPT-4o mini OpenAI	$0.15	$0.6
GPT-5.4 mini OpenAI	$0.75	$4.5
GPT-6.1 Sol OpenAI	$2	$10
Gemini 3.5 Flash-Lite Google	$0.3	$2.5
Gemini 3.8 Flash Google	$0.75	$3.75
Claude Haiku 4.5 Anthropic	$1	$5
Claude Sonnet 5.5 Anthropic	$2	$10
Claude Opus 5.5 Anthropic	$4	$20

LLM cost calculator questions

How is the cost worked out?: Cost per call is input tokens times the model's input price, plus output tokens times its output price, divided by a million. The month is calls a day times 30. It uses each vendor's list price, with no discounts, no prompt caching and no batch pricing.
Why is Jev so much cheaper?: Jev bills input tokens only, at $0.042 per million, and charges nothing for output, because it returns a typed answer (yes or no, a choice, a score) instead of text. A chat model bills for every token it writes. When your task is a decision, that difference is most of the bill.
Can Jev replace the other models here?: Only for decisions. It cannot write, summarize or answer in prose. Tick the box that says your calls need written text, and Jev leaves the comparison. For the jobs it does suit, see how it compares in the published runs on the Jev vs an LLM page.
Are the prices current?: They were copied from each vendor's pricing page on 2026-10-02, and each model name links to its source. Vendors change prices, and Google lists a rise for Gemini 3.8 Flash from 1 January 2027, so check the source before you commit a budget.
Why do the same tokens cost differently by model?: Tokenizers differ. The same text can be a different number of tokens on different models, and Anthropic says its newer tokenizer produces about 30% more tokens for the same text. The calculator takes your token counts at face value, so measure them on each model you compare.
Does this tool call a model or save anything?: No. It is arithmetic in your browser. The numbers live only in the page link, so you can share them.

What Jev costs on real work

The calculator uses list prices. For what people actually paid, with the volume and the bill they published, see Jev pricing and Jev vs an LLM.