13 model phổ biến (OpenAI, Anthropic, Google, DeepSeek, Meta, Mistral, Qwen). Chọn use case → tự động compute cost theo USD + VND/tháng.
Use case
Hỏi đáp ngắn · 200K tokens/tháng
Input/call800 tokens
Output/call200 tokens
Calls/ngày500
Tokens/tháng15.0M
Filter tier
Gemini 2.5 FlashGoogle1000K ctx
$1.80 · 45.900 đ
in: $0.075/Mout: $0.3/M
GPT-4o miniOpenAI128K ctx
$3.60 · 91.800 đ
in: $0.15/Mout: $0.6/M
DeepSeek V3.2DeepSeek128K ctx
$6.54 · 166.770 đ
in: $0.27/Mout: $1.1/M
Qwen 2.5 72BAlibaba128K ctx
$8.40 · 214.200 đ
in: $0.4/Mout: $1.2/M
Llama 3.3 70BMeta/Groq128K ctx
$9.45 · 240.975 đ
in: $0.59/Mout: $0.79/M
Claude 4.5 HaikuAnthropic200K ctx
$21.60 · 550.800 đ
in: $0.8/Mout: $4/M
Gemini 2.5 ProGoogle2000K ctx
$30.00 · 765.000 đ
in: $1.25/Mout: $5/M
Mistral Large 2Mistral128K ctx
$42.00 · 1.071.000 đ
in: $2/Mout: $6/M
GPT-4oOpenAI128K ctx
$60.00 · 1.530.000 đ
in: $2.5/Mout: $10/M
o1-miniOpenAI128K ctx
$72.00 · 1.836.000 đ
in: $3/Mout: $12/M
Claude 4.5 SonnetAnthropic200K ctx
$81.00 · 2.065.500 đ
in: $3/Mout: $15/M
GPT-5 (preview)OpenAI200K ctx
$120.00 · 3.060.000 đ
in: $5/Mout: $20/M
Claude 4.1 OpusAnthropic200K ctx
$405.00 · 10.327.500 đ
in: $15/Mout: $75/M
* Giá retail public API (USD per 1M tokens), không tính batch discount, prompt caching, hoặc volume tier. Lưu ý: prompt caching (Anthropic, OpenAI, Gemini) có thể giảm 50-90% input cost cho repeated context.