Know the LLM bill here. See what a finished build looks like there.
AgentHive Inc · Palm Coast · September 2026 list prices
Know the API bill before you buy the desk.
AI Pricing Calculators is the free LLM rate card from AgentHive Inc in Palm Coast. It prices a workload on the September 2026 list rates for GPT-5.6, Claude, Gemini, Grok, DeepSeek, and Mistral, with cache and batch where the provider lists them.
- GPT-5.6 Sol$4 / $20
- Claude Fable 5$10 / $50
- Gemini 3.1 Flash-Lite$0.25 / $1.50
Built in Palm Coast for funded teams of 30 to 200.
The rate card
Estimate the monthly API cost
Enter a workload. The table ranks the September 2026 list prices. Treat the number as a planning figure and confirm it on the provider’s page before you commit spend. The same prices are on each model page for anyone who does not run JavaScript.
3,000Requests / month
DeepSeek V4 Flash (off-peak) · $1.65/moLowest monthly
$103Spread versus highest
| DeepSeek V4 Flash (off-peak) | DeepSeek | $0.22 | $0.66 | 1M | $0.0006 | $1.65 |
| GPT-5.6 Luna | OpenAI | $0.20 | $1.20 | 400k | $0.0008 | $2.40 |
| Gemini 3.1 Flash-Lite | $0.25 | $1.50 | 1M | $0.001 | $3.00 | |
| DeepSeek V4 Flash (peak) | DeepSeek | $0.44 | $1.32 | 1M | $0.0011 | $3.30 |
| Mistral Large 3 | Mistral | $0.50 | $1.50 | 256k | $0.0013 | $3.75 |
| Gemini 3.5 Flash-Lite | $0.30 | $2.50 | 1M | $0.0015 | $4.65 | |
| DeepSeek V4 Pro (off-peak) | DeepSeek | $0.66 | $1.98 | 1M | $0.0016 | $4.95 |
| Grok 4.3 Under 200k prompt tokens |
xAI | $1.25 | $2.50 | 1M | $0.0025 | $7.50 |
| Gemini 3.8 Flash Paid promo through 31 Dec 2026 |
$0.75 | $3.75 | 1M | $0.0026 | $7.87 | |
| DeepSeek V4 Pro (peak) | DeepSeek | $1.32 | $3.96 | 1M | $0.0033 | $9.90 |
| Claude Haiku 4.5 | Anthropic | $1 | $5 | 200k | $0.0035 | $10.50 |
| Grok 4.5 Under 200k prompt tokens |
xAI | $2 | $6 | 500k | $0.005 | $15.00 |
| Grok 4.6 Under 200k prompt tokens |
xAI | $2 | $6 | 500k | $0.005 | $15.00 |
| Gemini 3.5 Flash | $1.50 | $9 | 1M | $0.006 | $18.00 | |
| Claude Sonnet 5 | Anthropic | $2 | $10 | 200k | $0.007 | $21.00 |
| Gemini 3.1 Pro At or under 200k prompt tokens |
$2 | $12 | 1M | $0.008 | $24.00 | |
| GPT-5.6 Terra | OpenAI | $2 | $12 | 400k | $0.008 | $24.00 |
| GPT-5.6 Sol Promo list price through at least 21 Nov 2026 |
OpenAI | $4 | $20 | 400k | $0.014 | $42.00 |
| Claude Opus 4.8 | Anthropic | $5 | $25 | 200k | $0.018 | $52.50 |
| Claude Fable 5 | Anthropic | $10 | $50 | 200k | $0.035 | $105 |
| GPT-6 Astra | OpenAI | $10 | $50 | 400k | $0.035 | $105 |
About 750 English words is 1,000 tokens. Output is often the larger line on the rate card, so long replies move the bill more than long prompts.
00 chars · 0 words
—Lowest / call
| Model | Input / 1M | Output / 1M | Cost / call | Cost × copies |
|---|---|---|---|---|
| Paste text to estimate tokens. | ||||
Blended is a 3:1 input-to-output mix, a fast way to rank models on mixed work. Cache and batch in this column follow the monthly tab.
| DeepSeek V4 Flash (off-peak) | DeepSeek | $0.22 | $0.66 | $0.007 | 1M | $0.33 | No |
| GPT-5.6 Luna | OpenAI | $0.20 | $1.20 | $0.02 | 400k | $0.45 | Yes |
| Gemini 3.1 Flash-Lite | $0.25 | $1.50 | $0.025 | 1M | $0.56 | Yes | |
| DeepSeek V4 Flash (peak) | DeepSeek | $0.44 | $1.32 | $0.014 | 1M | $0.66 | No |
| Mistral Large 3 | Mistral | $0.50 | $1.50 | $0.05 | 256k | $0.75 | Yes |
| Gemini 3.5 Flash-Lite | $0.30 | $2.50 | $0.03 | 1M | $0.85 | Yes | |
| DeepSeek V4 Pro (off-peak) | DeepSeek | $0.66 | $1.98 | $0.022 | 1M | $0.99 | No |
| Grok 4.3 | xAI | $1.25 | $2.50 | $0.20 | 1M | $1.56 | No |
| Gemini 3.8 Flash | $0.75 | $3.75 | $0.075 | 1M | $1.50 | Yes | |
| DeepSeek V4 Pro (peak) | DeepSeek | $1.32 | $3.96 | $0.044 | 1M | $1.98 | No |
| Claude Haiku 4.5 | Anthropic | $1 | $5 | $0.10 | 200k | $2 | Yes |
| Grok 4.5 | xAI | $2 | $6 | $0.30 | 500k | $3 | No |
| Grok 4.6 | xAI | $2 | $6 | $0.50 | 500k | $3 | No |
| Gemini 3.5 Flash | $1.50 | $9 | $0.15 | 1M | $3.38 | Yes | |
| Claude Sonnet 5 | Anthropic | $2 | $10 | $0.20 | 200k | $4 | Yes |
| Gemini 3.1 Pro | $2 | $12 | $0.20 | 1M | $4.50 | Yes | |
| GPT-5.6 Terra | OpenAI | $2 | $12 | $0.20 | 400k | $4.50 | Yes |
| GPT-5.6 Sol | OpenAI | $4 | $20 | $0.40 | 400k | $8 | Yes |
| Claude Opus 4.8 | Anthropic | $5 | $25 | $0.50 | 200k | $10 | Yes |
| Claude Fable 5 | Anthropic | $10 | $50 | $1 | 200k | $20 | Yes |
| GPT-6 Astra | OpenAI | $10 | $50 | $1 | 400k | $20 | Yes |
—Self-host / 1M tokens
—API blended / 1M
—Tokens / month where rent equals that API
—Tokens / month where API spend meets the seat
—Comparison model, 3:1 blend, no cache, no batch
First Deploy AI
AI work that looks as good as the rest of your company
For funded teams of 30 to 200 that already pay for AI. First Deploy builds the AI work your leadership will put its name on: finished, on brand, and running in your own accounts. Founding membership is the way in.
Prompt tutoring and a consulting tour are billed as consult time, below.
From the hive
Open comments
This calculator is the free door into the hive. The swarm products sit one click away.
Reed · AgentHive
Daniel Graham built the free rate card so operators can see cost before they talk.
Paid time
Consult, an operator in the room
Free 30-minute qualifier. Then paid time, or First Deploy when the build is clear. Book the free 30 on Calendly. Pay on Stripe when it is paid time.
Live apps
The rest of the board
These are the live products. Prices and URLs match the public pages.
Skip the meeting
Street $0 · Lights $19/mo · Marquee $99
Record and send without making the other person create an account first.
Proof the job happened
Solo $49/mo · Crew $99/mo
Photos, timestamps, and a record the crew can leave and the customer can see.
Tonight’s callback texts
$19.99 one-time
Paste today’s missed calls. Get every text to send before you lock up.
See the yearly card leak
$29.99 one-time
Paste last month’s card volume and what the processor kept. Effective rate, published bands, yearly leak versus 2.5%.
Built it? Get it found.
Pro $19.99 · Studio $29.99
Indexing desk: IndexNow, search-console steps, crawler files.
Lock down what bots can do
Field Kit $49 · Pro $149
Least privilege, audit, and deny-by-default tools for AI agents.
How the rate card works
What the calculator is modeling
APIs bill by the token, about 4 characters, or about three-quarters of an English word. You pay separately for input (prompt, system text, documents) and output (the reply). Almost every provider charges more for output.
Levers that move a bill
- Model choice. Claude Fable 5 lists at $10 / $50 per million. Gemini 3.1 Flash-Lite lists at $0.25 / $1.50.
- Prompt caching. Repeated prefixes are often billed near 10% of the input rate on OpenAI, Anthropic, and Google. The cache slider models that.
- Batch. Non-urgent work can run at half price where the provider lists a batch API. The monthly tab says so on the row.
- Self-hosting. A rented GPU costs the same whether it is busy. The self-host tab shows a floor, and it tells you when one box cannot reach the break-even volume.
A worked example
1,000 input tokens and 500 output tokens, on GPT-5.6 Sol at $4 in / $20 out per million: input is $0.004; output is $0.010. The call is $0.014. Output is three-quarters of that call. The full walk-through is on the rate card page.
Questions
Straight answers
Is the calculator free?
Yes. No signup and no paywall on this site. The paid offers are First Deploy AI, consult time, and the live apps listed on the page.
How current are the model prices?
They were read from official provider pages in September 2026: OpenAI, Anthropic, Google Gemini, xAI, DeepSeek, and Mistral. Providers change rates. Confirm before you buy volume.
What does GPT-5.6 Sol cost?
OpenAI’s short-context list price on this card is $4 per million input tokens and $20 per million output, with cached input at $0.40. Terra is $2 / $12. Luna is $0.20 / $1.20. Sol’s promotional framing on OpenAI’s page runs at least through 21 November 2026.
What is a blended rate?
Blended on this card is a 3:1 input-to-output mix: three parts input price plus one part output price, divided by four. It is a fast way to rank models on mixed work, not a quote for your exact prompt.
Does the batch discount apply to every model?
No. The monthly tab applies a 50% batch rate only where this card marks the model as listing a batch API. Grok and DeepSeek rows stay at list price when the batch box is checked.
What is First Deploy, exactly?
The done-for-you AI build for funded teams of 30 to 200 that already pay for AI. Finished, on brand, and running in your own accounts, so it looks as good as the rest of the company. Founding membership: https://firstdeploy.ai/founding.
How does consult billing work?
Book a free 30-minute qualifier. After that: $75 per 30 minutes, $150 per hour, or a 10-hour pack at $1,250 with $625 due up front. Stripe links are on this page and on firstdeploy.ai/consult.
Do you name customers?
No. We never name clients.