Back to AI models
Prices are estimates parsed from the catalog's public pricing summary. When the catalog has one blended token price, the calculator applies it to both input and output tokens. 16 model(s) were skipped because they do not have a usable dollar-per-token figure.
Rank Model Maker Input/mo Output/mo Total/mo Per request Relative cost
1 GPT-5 nano OpenAI ¤30 ¤8 ¤38 ¤0.0001
2 GPT-5 mini OpenAI ¤150 ¤38 ¤188 ¤0.0006
3 Claude Haiku 4.5 Anthropic ¤600 ¤150 ¤750 ¤0.0025
4 GPT-5.6 Luna OpenAI ¤600 ¤150 ¤750 ¤0.0025
5 GPT-5 OpenAI ¤750 ¤188 ¤938 ¤0.0031
6 Grok 4.3 xAI / SpaceXAI ¤750 ¤188 ¤938 ¤0.0031
7 Grok 4.5 xAI / SpaceXAI ¤1,200 ¤300 ¤1,500 ¤0.0050
8 GPT-5.6 Terra OpenAI ¤1,500 ¤375 ¤1,875 ¤0.0063
9 Claude Sonnet 5 Anthropic ¤1,800 ¤450 ¤2,250 ¤0.0075
10 Claude Opus 4.8 Anthropic ¤3,000 ¤750 ¤3,750 ¤0.0125
11 GPT-5.6 Sol OpenAI ¤3,000 ¤750 ¤3,750 ¤0.0125
12 Claude Fable 5 Anthropic ¤6,000 ¤1,500 ¤7,500 ¤0.0250
13 Claude Mythos 5 Anthropic ¤6,000 ¤1,500 ¤7,500 ¤0.0250

Self-Hosted Models

These models are not ranked as $0 because the cost moves to GPU infrastructure, engineering time, monitoring, and capacity planning.

Aya Expanse 32B Cohere For AI Command R7B Cohere DeepSeek-R1 DeepSeek DeepSeek-V3.2 DeepSeek DeepSeek-V4-Flash DeepSeek DeepSeek-V4-Pro DeepSeek Jamba2 3B AI21 Labs Llama 3.1 405B Instruct Meta Llama 3.2 3B Instruct Meta Llama 3.2 90B Vision Instruct Meta Llama 3.3 70B Instruct Meta Llama 4 Maverick Meta Llama 4 Scout Meta Llama-3.1-Nemotron-Ultra-253B-v1 NVIDIA Ministral 3 14B Mistral AI Ministral 3 3B Mistral AI Ministral 3 8B Mistral AI Mistral Large 3 Mistral AI Mistral Medium 3.5 Mistral AI Mistral Small 4 Mistral AI Nemotron 3 Ultra NVIDIA Phi-4 Microsoft Phi-4-mini-instruct Microsoft Phi-4-mini-reasoning Microsoft Phi-4-multimodal-instruct Microsoft Phi-4-reasoning Microsoft Qwen3-235B-A22B Alibaba Cloud / Qwen Qwen3-30B-A3B Alibaba Cloud / Qwen Qwen3-32B Alibaba Cloud / Qwen Qwen3-Coder-480B-A35B Alibaba Cloud / Qwen