On June 13, 2023, OpenAI announced price cuts for GPT-3.5-turbo (25% reduction to $0.0015/$0.002 per 1K input/output tokens) and text-embedding-ada-002 (75% reduction to $0.0001 per 1K tokens), attributed to increased system efficiency. The company also released function-calling capability for GPT-3.5-turbo and GPT-4, enabling more reliable structured outputs, and introduced a GPT-3.5-turbo variant with 16,000-token context window at 2x base price. CEO signaled that post-GPT-4 strategy is incremental model refinement rather than new frontier training. The pricing cuts indicate either margin headroom or deliberate market-share competition at the inference layer; the feature additions signal focus on developer usability over raw capability gains.
176 other tracked developments touching the same claims.