Work out what your AI feature actually costs
Fourteen free calculators and a price index covering 101 individually sourced prices across LLM APIs, embeddings, image and video generation, speech and GPU rental. Re-verified from vendor pricing pages every week. Everything runs in your browser — nothing you type is sent anywhere.
27 LLM price points 22 GPU rates 24 image & video models 15 speech models Verified 17 August 2026
Start with the question you have
LLM API costs
Price a feature and see what a request actually costs.
LLM API cost calculatorPrice an LLM feature per month across 27 models from OpenAI, Anthropic, Google, Mistral and DeepSeek, with caching and batch discounts included.
Token estimator and cost checkerPaste any text and estimate its token count, then see what it costs to send through 27 different LLM APIs. Runs entirely in your browser — nothing is uploaded.
AI chatbot cost calculatorPrice an AI chat feature including the history you resend every turn. Compare full history, sliding window and rolling summary side by side.
Context window cost calculatorCheck whether your document fits a model's context window, what a full-context call costs, and at what query volume retrieval becomes cheaper than stuffing.
Cut your bill
What caching, batching and model choice save you.
Prompt caching savings calculatorFind out how much prompt caching actually saves on your workload, and the hit rate you need before the cache-write premium pays for itself.
Batch API savings calculatorSee exactly what the 50% batch discount is worth on your workload, and how much of your traffic can tolerate the asynchronous turnaround.
Model switch savings calculatorPrice your current workload on your current model, then on every alternative, and see exactly what a migration would save or cost per year.
Fine-tuning vs prompting cost calculatorFine-tuning buys a shorter prompt. Work out how many requests it takes for those saved tokens to repay the training run and the higher inference rate.
Build vs buy
Compare running it yourself against paying per token.
Self-hosting vs API break-even calculatorCompare renting a GPU to serve an open-weight model against paying per token, including utilisation, redundancy and the ops time nobody budgets for.
RAG pipeline cost calculatorPrice a retrieval-augmented generation pipeline end to end: initial embedding, re-embedding on updates, vector storage and the generation cost of every query.
AI subscription vs API cost calculatorCompare paying per seat for AI assistants against building on the API directly, and find the team size where the answer flips.
Media, speech & ROI
Images, video, audio and whether it all pays off.
AI image and video generation cost calculatorPrice a generative image or video pipeline across Imagen, FLUX, Stable Diffusion, Veo, Kling, Luma and Runway — including the retries nobody budgets for.
Speech-to-text and text-to-speech cost calculatorPrice transcription and synthetic voice at your volume across OpenAI, Deepgram, AssemblyAI, Google and ElevenLabs, with the real cost per hour of audio.
AI ROI calculatorTurn hours saved into money, subtract what the tooling costs, and get a payback period with a sensitivity range instead of a single optimistic number.
Cheapest models right now
Blended cost of one million tokens at a 4:1 input-to-output ratio — a realistic shape for summarisation and extraction work. Full detail on the price index.
| Model | Input $/Mtok | Output $/Mtok | Blended 4:1 |
|---|---|---|---|
| Mistral AI Ministral 3 (3B) | $0.10 | $0.10 | $0.100 |
| Google Gemini 2.5 Flash-Lite | $0.10 | $0.40 | $0.160 |
| DeepSeek DeepSeek V4 Flash | $0.22 | $0.28 | $0.232 |
| Meta via Fireworks AI Llama 4 Scout Instruct | $0.15 | $0.60 | $0.240 |
| Mistral AI Mistral Small 4 | $0.15 | $0.60 | $0.240 |
| Qwen via Together AI Qwen3 235B A22B Instruct | $0.20 | $0.60 | $0.280 |
| Qwen via Together AI Qwen2.5 7B Instruct Turbo | $0.30 | $0.30 | $0.300 |
| Cohere Command-light | $0.30 | $0.60 | $0.360 |
| OpenAI GPT-5.6 Luna | $0.20 | $1.20 | $0.400 |
| Mistral AI Codestral | $0.30 | $0.90 | $0.420 |
Latest price movements
| Date | Item | Was | Now | Change |
|---|---|---|---|---|
| 2026-08-17 | DeepSeek DeepSeek V4 Flash | $0.14 | $0.22 | +57.1% |
| 2026-08-17 | DeepSeek DeepSeek V4 Pro | $0.435 | $0.66 | +51.7% |
| 2026-08-16 | ElevenLabs Scribe v2 (Pro rate) | $0.0545 | $0.054 | -0.9% |
Understand the numbers
How LLM token pricing actually worksWhat a token is, why input and output are billed separately, how cached and batch tiers work, and a full worked example priced on two models.
Why output tokens cost more than input tokensPrefill runs in parallel, decode runs one token at a time against memory bandwidth. That mechanical difference is what the 5:1 price ratio is pricing.
Eleven ways to cut your LLM bill without switching modelOrdered by how much they actually return: context surgery, caching, batching, routing, output caps and the instrumentation that finds the rest.
Prompt caching explained, with the real break-evenThe byte-identical prefix rule, write premiums, cache lifetime, and the hit rate below which caching costs more than it saves on Claude Sonnet 5.
Batch API processing: when the 50% discount is worth itA feature-by-feature audit of which AI workloads tolerate a 24-hour batch window, priced at real rates, plus what vendors promise and what fails.
How to choose a model tier without guessingBuild a 50-example eval set, run the price ladder from cheapest upward, and stop at the first model that passes. With the full cost spread worked out.
RAG vs long context: run the crossover mathsStuffing a 239,000-token corpus costs $0.072 a query on Gemini 2.5 Flash. Retrieval costs $0.0012. Here is where the crossover actually falls, and why.
The real cost of self-hosting an open-weight modelA two-H100 deployment costs $4,803 a month whether you use it or not. Worked against Llama 3.3 70B on Together, with the overheads nobody budgets for.
The hidden costs of AI that never reach your price cardAn API dashboard reading $1,120 a month for a support feature that really costs $12,400 to run, with retries, evals, guardrails and review sized.
Estimating tokens before launch, with no traffic yetRules of thumb, a representative sample, why the 95th percentile beats the median, a worked support-summarisation estimate and four costly omissions.
LLM cost forecasting for a bill that has not happenedThree scenarios instead of one, unit cost per business event, why AI unit cost stays flat with scale, and the four things that cause a step change.
Embedding model selection, quality first and price lastEmbedding a 25,000-document corpus costs under two dollars, so choose on retrieval quality and dimensions. What actually costs money is downstream.