OpenAI Discontinues Fine-Tuning and Legacy Models, Introduces GPT-4.5 and New o3-mini Pricing for Enhanced Accessibility

Fine-tuning for gpt-3.5-turbo, davinci-002, and babbage-002 models removed

Legacy models section removed (chatgpt-4o-latest, gpt-4-turbo, gpt-4, gpt-4-32k, gpt-3.5-turbo variants, davinci-002, babbage-002)

Small reasoning model pricing decreased 63%: o1-mini ($3 input / $12 output) replaced by o3-mini ($1.10 input / $4.40 output) per 1M tokens

o1 reasoning models context length increased from 128K to 200K tokens

Realtime API added for low-latency multimodal experiences including speech-to-speech

GPT-4.5 model added for creative tasks and agentic planning at $75 input / $150 output per 1M tokens (research preview)

Assistants API tool pricing removed (Code Interpreter $0.03/session, File Search $0.10/GB/day)

OpenAI o1 frontier reasoning model added with tools, Structured Outputs, and vision support (200k context) at $15 input / $60 output per 1M tokens

o1-mini model discontinued (replaced by o3-mini)

Cached input pricing introduced offering 50% discount on cached inputs across multiple models

OpenAI o3-mini small reasoning model added at $1.10 input / $4.40 output per 1M tokens

Image generation models section removed (DALL·E 3 Standard, HD, DALL·E 2)

Embedding models section removed (text-embedding-3-small, text-embedding-3-large, ada v2)

o1-preview model discontinued (replaced by o1)

Audio models section removed (Whisper, TTS, TTS HD)