KI Modell Text,Hersteller,Input Costs / 1M tokens,Cached Input / 1M tokens,Output Costs / 1M tokens,Input Toke Limit,Output Token Limit,Batch API,Hoster,Training Data Cut-off,"Kosten pro 1M Input 0,5M Output"
o1,OpenAI,15,"7,5",60,200000,100000,50% for 24 hours,"OpenAI, Azure",Oktober 2023,45
gpt 4.1,OpenAI,2,"0,5",8,1000000,32768,50% for 24 hours,"OpenAI, Azure",Juni 2024,6
GPT-4o,OpenAI,"2,5","1,25",10,128000,16384,50% for 24 hours,"OpenAI, Azure",Oktober 2023,"7,5"
gpt 4.1-mini,OpenAI,"0,4","0,1","1,6",1000000,32768,50% for 24 hours,"OpenAI, Azure",Juni 2024,"1,2"
gpt 4.1-nano,OpenAI,"0,1","0,025","0,4",1000000,32768,50% for 24 hours,"OpenAI, Azure",Juni 2024,"0,3"
o3-mini,OpenAI,"1,1","0,55","4,4",200000,100000,50% for 24 hours,"OpenAI, Azure",Oktober 2023,"3,3"
Claude Sonnet 3.7,Antrophic,3,Up to 90%,15,200000,8192,50% for 24 hours,"Google, Amazon, Antrophic",April 2024,"10,5"
Claude Haiku 3.5,Antrophic,"0,8",Up to 90%,4,200000,8192,50% for 24 hours,"Google, Amazon, Antrophic",Juli 2024,"2,8"
Deepseek R1,Deepseek,"0,55","0,14","2,19",64000,8192,,Deepseek,Juli 2024,"1,645"
Deepseek V3,Deepseek,"0,27","0,07","1,1",64000,8192,,Deepseek,Dezember 2024,"0,82"
GPT-4o mini,OpenAI,"0,15","0,075","0,6",128000,16384,50% for 24 hours,"OpenAI, Azure",Oktober 2023,"0,45"
Gemini 2.5 Pro,,"1,25",,10,1.048.576,65.536,,Google,Januar 2025,"6,25"
Gemini 2.0 Flash,Google,"0,1","0,025","0,4",1.048.576,8192,,Google,Juni 2024,"0,3"
Gemini 2.0 Flash-Lite,Google,"0,075","0,01875","0,3",1.048.576,8192,,Google,Juni 2024,"0,225"