LLM API cost calculator
Enter your prompt and response sizes and how many requests you make, and compare what 34 models cost per request and per month — side by side, from each model's published per-token prices.
Estimate your monthly bill
Your inputs are saved in the page address, so you can share the exact comparison as a link.
Prompt, system instructions and any chat history sent
The model's reply (include reasoning tokens if billed)
= 30,000 requests a month (30-day month).
| On MultiverseAI credits | |||
|---|---|---|---|
| GPT-5 NanoOpenAI | $0.00025 | $7.50 | $11.251,125 credits |
| Llama 3.3 70BMeta | $0.00026 | $7.80 | $11.701,170 credits |
| GPT-6 LunaOpenAI | $0.00035 | $10.50 | $15.751,575 credits |
| Claude Haiku 5.5Anthropic | $0.00035 | $10.50 | $15.751,575 credits |
| GPT-4o MiniOpenAI | $0.00045 | $13.50 | $20.252,025 credits |
| GPT-OSS 120BOpenAI | $0.00045 | $13.50 | $20.252,025 credits |
| Llama 4 MaverickMeta | $0.00051 | $15.41 | $23.122,312 credits |
| DeepSeek V3DeepSeek | $0.00077 | $23.15 | $34.733,473 credits |
| GPT-5 MiniOpenAI | $0.0013 | $37.50 | $56.255,625 credits |
| Gemini 2.5 FlashGoogle | $0.0015 | $46.50 | $69.756,975 credits |
| DeepSeek R1DeepSeek | $0.0016 | $47.25 | $70.887,088 credits |
| Gemini 3 FlashGoogle | $0.002 | $60.00 | $90.009,000 credits |
| Gemini 3.8 FlashGoogle | $0.0029 | $86.63 | $13012,994 credits |
| Claude Haiku 4.5Anthropic | $0.0035 | $105 | $15815,750 credits |
| Grok 4.7xAI | $0.005 | $150 | $22522,500 credits |
| Mistral LargeMistral | $0.005 | $150 | $22522,500 credits |
| o3OpenAI | $0.006 | $180 | $27027,000 credits |
| GPT-5.1OpenAI | $0.0063 | $188 | $28128,125 credits |
| GPT-5OpenAI | $0.0063 | $188 | $28128,125 credits |
| Gemini 2.5 ProGoogle | $0.0063 | $188 | $28128,125 credits |
| SonarPerplexity | $0.0065 | $195 | $29329,250 credits |
| GPT-6.1 SolOpenAI | $0.007 | $210 | $31531,500 credits |
| Claude Sonnet 5.5Anthropic | $0.007 | $210 | $31531,500 credits |
| GPT-4oOpenAI | $0.0075 | $225 | $33833,750 credits |
| Gemini 3.1 ProGoogle | $0.008 | $240 | $36036,000 credits |
| Claude 4.6 SonnetAnthropic | $0.011 | $315 | $47347,250 credits |
| Claude Sonnet 4.5Anthropic | $0.011 | $315 | $47347,250 credits |
| Sonar Deep ResearchPerplexity | $0.011 | $330 | $49549,500 credits |
| Sonar Reasoning ProPerplexity | $0.011 | $330 | $49549,500 credits |
| Claude Opus 5.5Anthropic | $0.014 | $420 | $63063,000 credits |
| Sonar ProPerplexity | $0.015 | $465 | $69869,750 credits |
| Claude Opus 4.6Anthropic | $0.018 | $525 | $78878,750 credits |
| Claude Fable 5.1Anthropic | $0.035 | $1,050 | $1,575157,500 credits |
| GPT-5 ProOpenAI | $0.075 | $2,250 | $3,375337,500 credits |
Provider prices from OpenRouter, checked on November 2, 2026. The MultiverseAI column is provider cost + 50% in credits (1 credit = $0.01) — the same rates the platform bills; with your own API keys it's 0 credits. Perplexity's Sonar models also charge a search fee per request, on top of tokens: OpenRouter lists $0.005, and the per-request figures here include it. Perplexity's own fee rises with the search context size, and Sonar Deep Research additionally bills citation tokens, reasoning tokens and each search query it runs, which aren't included. On MultiverseAI, Sonar calls are billed at the cost Perplexity reports for the request, plus 50%.
LLM API prices per million tokens
Standard input and output prices in USD per 1M tokens, from OpenRouter's public model list, checked on November 2, 2026. Prices change; check the source before you budget.
| Model | Provider | Input / 1M tokens | Output / 1M tokens | Context window | MultiverseAI credits / 1M in · out |
|---|---|---|---|---|---|
| Claude Haiku 5.5$0.50 · $2.50 above 100K input | Anthropic | $0.10 | $0.50 | 1M | 15 · 75 |
| Claude Haiku 4.5 | Anthropic | $1.00 | $5.00 | 200K | 150 · 750 |
| Claude Sonnet 5.5 | Anthropic | $2.00 | $10.00 | 1M | 300 · 1,500 |
| Claude 4.6 Sonnet | Anthropic | $3.00 | $15.00 | 1M | 450 · 2,250 |
| Claude Sonnet 4.5$6.00 · $22.50 above 200K input | Anthropic | $3.00 | $15.00 | 1M | 450 · 2,250 |
| Claude Opus 5.5 | Anthropic | $4.00 | $20.00 | 1M | 600 · 3,000 |
| Claude Opus 4.6 | Anthropic | $5.00 | $25.00 | 1M | 750 · 3,750 |
| Claude Fable 5.1 | Anthropic | $10.00 | $50.00 | 1M | 1,500 · 7,500 |
| DeepSeek V3 | DeepSeek | $0.2574 | $1.0287 | 164K | 39 · 154 |
| DeepSeek R1 | DeepSeek | $0.50 | $2.15 | 164K | 75 · 322 |
| Gemini 2.5 Flash | $0.30 | $2.50 | 1.05M | 45 · 375 | |
| Gemini 3 Flash | $0.50 | $3.00 | 1.05M | 75 · 450 | |
| Gemini 3.8 Flash | $0.825 | $4.125 | 1.05M | 124 · 619 | |
| Gemini 2.5 Pro$2.50 · $15.00 above 200K input | $1.25 | $10.00 | 1.05M | 188 · 1,500 | |
| Gemini 3.1 Pro$4.00 · $18.00 above 200K input | $2.00 | $12.00 | 1.05M | 300 · 1,800 | |
| Llama 3.3 70B | Meta | $0.10 | $0.32 | 131K | 15 · 48 |
| Llama 4 Maverick | Meta | $0.1875 | $0.6525 | 524K | 28 · 98 |
| Mistral Large | Mistral | $2.00 | $6.00 | 128K | 300 · 900 |
| GPT-5 Nano | OpenAI | $0.05 | $0.40 | 400K | 8 · 60 |
| GPT-6 Luna$0.20 · $0.75 above 272K input | OpenAI | $0.10 | $0.50 | 1.05M | 15 · 75 |
| GPT-4o Mini | OpenAI | $0.15 | $0.60 | 128K | 22 · 90 |
| GPT-OSS 120B | OpenAI | $0.15 | $0.60 | 131K | 22 · 90 |
| GPT-5 Mini | OpenAI | $0.25 | $2.00 | 400K | 38 · 300 |
| GPT-5.1 | OpenAI | $1.25 | $10.00 | 400K | 188 · 1,500 |
| GPT-5 | OpenAI | $1.25 | $10.00 | 400K | 188 · 1,500 |
| GPT-6.1 Sol$4.00 · $15.00 above 272K input | OpenAI | $2.00 | $10.00 | 1.05M | 300 · 1,500 |
| o3 | OpenAI | $2.00 | $8.00 | 200K | 300 · 1,200 |
| GPT-4o | OpenAI | $2.50 | $10.00 | 128K | 375 · 1,500 |
| GPT-5 Pro | OpenAI | $15.00 | $120.00 | 400K | 2,250 · 18,000 |
| Sonar+ $0.005 search fee per request | Perplexity | $1.00 | $1.00 | 127K | 150 · 150 |
| Sonar Deep Research+ $0.005 search fee per request | Perplexity | $2.00 | $8.00 | 128K | 300 · 1,200 |
| Sonar Reasoning Pro+ $0.005 search fee per request | Perplexity | $2.00 | $8.00 | 128K | 300 · 1,200 |
| Sonar Pro+ $0.005 search fee per request | Perplexity | $3.00 | $15.00 | 200K | 450 · 2,250 |
| Grok 4.7$4.00 · $12.00 above 200K input | xAI | $2.00 | $6.00 | 500K | 300 · 900 |
How the numbers are worked out
Cost per request
Input tokens × input price ÷ 1,000,000, plus output tokens × output price ÷ 1,000,000. Monthly cost is that times your requests a month; requests per day are multiplied by 30.
Words and characters
Converted with the usual rule of thumb for English: about 4 characters, or 0.75 of a word, per token. Every model family has its own tokenizer, so these are estimates — count real tokens when you can.
On MultiverseAI
Platform usage is billed at provider cost × 1.5 in credits of $0.01 — a $0.002 request is 0.30 credits. That's more than going to the provider directly; with your own API keys there's no credit charge. See pricing.
LLM API pricing: common questions
How is LLM API cost calculated?
Providers charge separately for input (prompt) tokens and output (completion) tokens, usually quoted per million tokens. Cost per request = input tokens × input price ÷ 1,000,000 + output tokens × output price ÷ 1,000,000. Multiply by your request volume for a monthly figure; this calculator counts a month as 30 days.
How many tokens is a word?
For English text, one token is roughly 4 characters or about 0.75 of a word, so 1,000 words is around 1,330 tokens. Code, non-English text and unusual formatting usually take more tokens. Each model family uses its own tokenizer, so treat word and character conversions as estimates.
Where do these prices come from?
From OpenRouter's public model list (https://openrouter.ai/api/v1/models), which publishes a per-token input and output price for every model it serves; for OpenAI, Anthropic, Google and Perplexity models that is the provider's own list price. The snapshot on this page was checked on 2026-11-02. Prices change — check the source before you commit to a budget.
Why do output tokens cost more than input tokens?
Generating each output token takes a full pass through the model, while input tokens are processed together in parallel. Providers price that difference in: across the models on this page, output tokens cost between 1× and 8.3× the input price. Long answers therefore drive cost much more than long prompts.
What does the same usage cost on MultiverseAI?
MultiverseAI bills AI usage in credits at the provider's cost plus 50%, where 1 credit = $0.01. A request that costs $0.002 at the provider costs 0.3 credits ($0.003). Fractions of a credit carry over until they add up to a whole credit. That is more than calling the provider yourself — the margin pays for the platform: agents, tools, workflows and one account across many models.
Can I use my own API keys on MultiverseAI instead?
Yes. Calls made with your own provider API keys are not charged credits — you pay your provider directly at their rates.
Does this include caching, batch or reasoning discounts and surcharges?
No. The calculator uses each model's standard input and output price (and its higher long-prompt price where one applies). Prompt caching and batch processing can lower the bill, and reasoning models may bill hidden reasoning tokens as output — include those in your output estimate if your model uses them. Perplexity's Sonar models also charge a search fee per request, on top of tokens: OpenRouter lists $0.005, and the per-request figures here include it. Perplexity's own fee rises with the search context size, and Sonar Deep Research additionally bills citation tokens, reasoning tokens and each search query it runs, which aren't included. On MultiverseAI, Sonar calls are billed at the cost Perplexity reports for the request, plus 50%.
Building on these models? The MultiverseAI API serves all of them behind one OpenAI-compatible endpoint.
Every model, one account
Chat with these models, build agents and run workflows on MultiverseAI. Every account starts with 100 free credits.