LLMCostLab

LLM API pricing, verified and turned into real numbers

Every model below is priced from the provider's own published rate card. The tables are useful; the calculators are the point — a $2 and a $10 model are not 5x apart once caching and your input:output ratio are accounted for.

Prices verified 2026-08-19 · 22 models · 3 providers
advertisement
Cheapest blended rate
Gemini 2.5 Flash-Lite
$0.175 /1M at 3:1 in:out
Priciest blended rate
GPT-5.6 Cyber
$28.12 /1M at 3:1 in:out
Spread, top to bottom
161x
before caching

All models

Blended rate assumes 3 input tokens per 1 output token, the rough shape of most production traffic. Sorted cheapest first.

ModelProviderInput /1MOutput /1MCached in /1MContextBlended /1M
Gemini 2.5 Flash-LiteGoogle$0.100$0.4001M$0.175
GPT-5.6 LunaOpenAI$0.100$0.600$0.010400k$0.225
Gemini 3.1 Flash-LiteGoogle$0.250$1.5001M$0.562
Gemini 3.5 Flash-LiteGoogle$0.300$2.5001M$0.850
Gemini 2.5 FlashGoogle$0.300$2.5001M$0.850
Gemini 3.7 FlashGoogle$0.750$3.7501M$1.500
Gemini 3.6 FlashGoogle$0.750$3.7501M$1.500
Claude Haiku 4.5Anthropic$1.000$5.000$0.100200k$2.000
GPT-5.6 TerraOpenAI$1.000$6.000$0.100400k$2.250
Gemini 3.5 FlashGoogle$1.500$9.0001M$3.375
Gemini 2.5 ProGoogle$1.250$10.001M$3.438
Claude Sonnet 5Anthropic$2.000$10.00$0.200200k$4.000
Gemini 3.1 ProGoogle$2.000$12.001M$4.500
GPT-5.3 CodexOpenAI$1.750$14.00$0.175400k$4.812
GPT-5.6 SolOpenAI$2.500$15.00$0.250400k$5.625
Claude Sonnet 4.6Anthropic$3.000$15.00$0.300200k$6.000
Claude Opus 5Anthropic$5.000$25.00$0.500200k$10.00
Claude Opus 4.5Anthropic$5.000$25.00$0.500200k$10.00
GPT chat-latestOpenAI$5.000$30.00$0.500400k$11.25
Claude Fable 5Anthropic$10.00$50.00$1.000200k$20.00
Claude Mythos 5Anthropic$10.00$50.00$1.000200k$20.00
GPT-5.6 CyberOpenAI$12.50$75.00$1.250400k$28.12

Price your own workload

Cost calculator

Enter your own numbers, or start from a workload preset. Costs are monthly and update as you type.

Cheapest model
Cheapest → priciest spread
ModelIn /1MOut /1MMonthlyvs best
advertisement

Workload calculators

Pre-built for the eight workloads that account for most production LLM spend. Each ranks every model on that workload's actual token shape.

Customer support chatbotRAG document Q&ACoding agentBulk summarizationStructured data extractionLong-form content generationEmail triage agentTranslation pipeline

Head-to-head comparisons

GPT-5.6 Sol vs GPT-5.6 TerraGPT-5.6 Sol vs GPT-5.6 LunaGPT-5.6 Sol vs GPT-5.3 CodexGPT-5.6 Sol vs Claude Opus 5GPT-5.6 Sol vs Claude Sonnet 5GPT-5.6 Sol vs Claude Haiku 4.5GPT-5.6 Sol vs Gemini 3.7 FlashGPT-5.6 Sol vs Gemini 3.1 ProGPT-5.6 Sol vs Gemini 2.5 Flash-LiteGPT-5.6 Terra vs GPT-5.6 LunaGPT-5.6 Terra vs GPT-5.3 CodexGPT-5.6 Terra vs Claude Opus 5GPT-5.6 Terra vs Claude Sonnet 5GPT-5.6 Terra vs Claude Haiku 4.5GPT-5.6 Terra vs Gemini 3.7 FlashGPT-5.6 Terra vs Gemini 3.1 ProGPT-5.6 Terra vs Gemini 2.5 Flash-LiteGPT-5.6 Luna vs GPT-5.3 CodexGPT-5.6 Luna vs Claude Opus 5GPT-5.6 Luna vs Claude Sonnet 5GPT-5.6 Luna vs Claude Haiku 4.5GPT-5.6 Luna vs Gemini 3.7 FlashGPT-5.6 Luna vs Gemini 3.1 ProGPT-5.6 Luna vs Gemini 2.5 Flash-LiteGPT-5.3 Codex vs Claude Opus 5GPT-5.3 Codex vs Claude Sonnet 5GPT-5.3 Codex vs Claude Haiku 4.5GPT-5.3 Codex vs Gemini 3.7 FlashGPT-5.3 Codex vs Gemini 3.1 ProGPT-5.3 Codex vs Gemini 2.5 Flash-LiteClaude Opus 5 vs Claude Sonnet 5Claude Opus 5 vs Claude Haiku 4.5Claude Opus 5 vs Gemini 3.7 FlashClaude Opus 5 vs Gemini 3.1 ProClaude Opus 5 vs Gemini 2.5 Flash-LiteClaude Sonnet 5 vs Claude Haiku 4.5Claude Sonnet 5 vs Gemini 3.7 FlashClaude Sonnet 5 vs Gemini 3.1 ProClaude Sonnet 5 vs Gemini 2.5 Flash-LiteClaude Haiku 4.5 vs Gemini 3.7 FlashClaude Haiku 4.5 vs Gemini 3.1 ProClaude Haiku 4.5 vs Gemini 2.5 Flash-LiteGemini 3.7 Flash vs Gemini 3.1 ProGemini 3.7 Flash vs Gemini 2.5 Flash-LiteGemini 3.1 Pro vs Gemini 2.5 Flash-Lite