Free tool

LLM pricing comparison

Every major model, per-million-token pricing, context window, modalities — sortable and bookmarkable.

← Back to free tools
Notes
Llama 3.1 8B (Groq)Groq$0.050$0.08020.0M128KtextLightning-fast on Groq's LPU hardware.
Gemini 1.5 FlashGoogle$0.075$0.3013.3M1Mtextvisionaudio
Gemini 2.0 FlashGoogle$0.10$0.4010.0M1Mtextvisionaudio
GPT-4o miniOpenAI$0.15$0.606.7M128Ktextvision
Mistral Small 3Mistral$0.20$0.605.0M32Ktext
DeepSeek V3DeepSeek$0.27$1.103.7M128KtextOff-peak (UTC 16:30–00:30) is 50% cheaper.
GPT-4.1 miniOpenAI$0.40$1.602.5M1Mtextvision
DeepSeek R1DeepSeek$0.55$2.191.8M128KtextReasoning model — outputs include chain-of-thought.
Llama 3.3 70B (Groq)Groq$0.59$0.791.7M128KtextOpen-weight model. Pricing + speed depend on host.
Claude Haiku 4.5Anthropic$0.80$41.3M200Ktextvision
o3-miniOpenAI$1.10$4.40909K200Ktext
Gemini 1.5 ProGoogle$1.25$5800K2Mtextvisionaudio2M context — the largest of any production model.
GPT-4.1OpenAI$2$8500K1Mtextvision
Grok-2xAI$2$10500K128Ktextvision
Mistral Large 2Mistral$2$6500K128Ktextvision
GPT-4oOpenAI$2.50$10400K128Ktextvision
Claude Sonnet 4.6Anthropic$3$15333K200Ktextvision
Claude 3.5 SonnetAnthropic$3$15333K200KtextvisionOlder but still popular for cost-stability reasons.
Grok-3xAI$3$15333K1Mtextvision
o1OpenAI$15$6067K200KtextReasoning tokens billed but hidden from output.
Claude Opus 4.7Anthropic$15$7567K1MtextvisionCache write 1.25× input price. Extended thinking optional. JSON via prefill or tool-use trick.

Last verified · 21 models shown · Data also available as JSON.

Source: provider pricing pages, May 2026. Prices can change — verify on each provider’s site for production use.

Want to track these costs on your actual traffic? Try Tokenwise — one line of code, $19/mo.