API pricing

LLM API Pricing by Provider

Calcaas tracks per-million-token pricing for 3,086 large language models across 45 providers, including OpenAI, Anthropic, Google Gemini, Vertex AI, AWS Bedrock and Azure OpenAI. Every provider page lists input cost, output cost, context window and capability support, current as of October 12, 2026.

Last updated October 12, 2026 · synced weekly from the upstream model catalog

Google Vertex AI pricing — Claude, Gemini, Llama, Mistral, DeepSeek and Qwen models on one page (132 models).

All providers

OpenRouter

353 models · from $0.030/1M output

Fireworks AI

322 models · from $0.0010/1M output

Vercel AI Gateway

256 models · from $0.030/1M output

Novita

209 models · from —/1M output

DeepInfra

206 models · from $0.020/1M output

Azure OpenAI

189 models · from —/1M output

AWS Bedrock

131 models · from $0.040/1M output

Together AI

111 models · from —/1M output

Azure AI

109 models · from —/1M output

OpenAI

106 models · from $0.140/1M output

Mistral AI

89 models · from $0.040/1M output

Databricks

81 models · from $0.200/1M output

Gemini on Vertex AI

78 models · from $0.250/1M output

Hugging Face

78 models · from $0.060/1M output

Nebius

76 models · from $0.030/1M output

Alibaba (DashScope)

69 models · from $0.035/1M output

Google Vertex AI

51 models · from $0.040/1M output

xAI (Grok)

43 models · from $0.500/1M output

Replicate

40 models · from $0.250/1M output

Qwen (Alibaba)

39 models · from $0.090/1M output

GitHub Copilot

35 models · from $0.500/1M output

Oracle Cloud

35 models · from $0.090/1M output

IBM watsonx

30 models · from $0.100/1M output

Anthropic

28 models · from $0.500/1M output

DeepSeek

24 models · from $0.100/1M output

Claude on Vertex AI

22 models · from $0.500/1M output

Perplexity

22 models · from $0.200/1M output

Lambda

20 models · from $0.025/1M output

Mistral (Vertex)

19 models · from $0.150/1M output

OVHcloud

19 models · from $0.100/1M output

Snowflake

19 models · from $0.240/1M output

Zhipu AI

18 models · from $0.100/1M output

Groq

16 models · from $0.030/1M output

Hyperbolic

16 models · from $0.200/1M output

Meta Llama

16 models · from $0.020/1M output

Cohere

15 models · from $0.150/1M output

FriendliAI

14 models · from $0.400/1M output

Nscale

14 models · from $0.025/1M output

Anyscale

12 models · from $0.150/1M output

Moonshot

12 models · from $2.00/1M output

MiniMax

11 models · from $1.02/1M output

AI21 Labs

9 models · from $0.400/1M output

Amazon Bedrock

8 models · from $0.140/1M output

Community

8 models · from $0.060/1M output

SambaNova

8 models · from $0.590/1M output