Free tool to estimate and compare monthly LLM API costs across GPT-4o, Claude 3.5 Sonnet, Gemini 1.5 Pro, and more. Adjust token usage and daily request volume — all calculations are client-side.
$0.075/M in • $0.3/M out • Ultra-fast & cheap.
$0.15/M in • $0.6/M out • Fast & very cheap.
$0.59/M in • $0.79/M out • Open-source via Groq.
$0.8/M in • $4/M out • Fast & affordable.
$1.25/M in • $5/M out • Long context (up to 2M).
$2.5/M in • $10/M out • Flagship multimodal model.
$3/M in • $12/M out • Reasoning model.
$3/M in • $15/M out • Best balanced model.
$10/M in • $30/M out • Previous flagship.
$15/M in • $75/M out • High intelligence.
Generating text is more compute-intensive than reading it. Output tokens require sequential autoregressive inference where each token depends on all previous ones, making output generation 3–5× more expensive than processing input tokens.
GPT-4o costs $2.50/1M input and $10/1M output tokens. Claude 3.5 Sonnet costs $3/1M input and $15/1M output. For read-heavy workloads like RAG, GPT-4o is cheaper on input. For balanced chat, GPT-4o wins on output too.
A token is the basic unit of text that a large language model reads and generates. On average, one token ≈ 4 characters or ¾ of an English word. A 1,000-word document is roughly 1,333 tokens. API pricing is quoted per million tokens (per 1M).
Six proven strategies: (1) Route simple tasks to cheaper models like GPT-4o mini or Claude 3.5 Haiku. (2) Cache frequent responses. (3) Shorten system prompts. (4) Set a lower max_tokens. (5) Batch requests where supported. (6) Use open-source models via Groq or Together AI.
Yes! Open-source models like Llama 3, Mistral, Gemma, and Phi-3 can be self-hosted for free (you pay only for compute). Services like Groq, Together AI, and Google AI Studio offer generous free tiers for supported open models.
Compare an illustrative monthly API budget across model presets.