SAN FRANCISCO – OpenAI, the leading artificial intelligence research company, announced a series of major upgrades and reduced pricing for its AI platform and tools during its annual DevDay event today. The announcements promise more advanced capabilities and greater accessibility for developers building AI-powered products and services.
New GPT-4 Turbo Model The highlight of the event was the introduction of GPT-4 Turbo, the next generation of OpenAI’s powerful natural language model. GPT-4 Turbo features increased capabilities and a massive 128,000 token context window – allowing it to ingest the equivalent of over 300 pages of text at once.
OpenAI also optimized GPT-4 Turbo’s performance, enabling a 3x reduction in input token pricing and a 2x reduction in output token pricing compared to the previous GPT-4 model. GPT-4 Turbo will be rolled out to all paying developers over the coming weeks.
Updated GPT-3.5 Turbo and Assistants API In addition to GPT-4, OpenAI released an updated version of GPT-3.5 Turbo with improved instruction following and support for new features like JSON mode. The company also launched a new Assistants API to make it easier for developers to build AI assistants with persistent memory and the ability to leverage tools like code execution, information retrieval and more.
Multimodal Capabilities Added Several new modalities were added to OpenAI’s platform, including vision support for GPT-4 to process images, text-to-speech to convert text into natural sounding voices, and integration of the DALL-E 3 image generation model.
Lower Prices Across Platform With optimizations in model training, OpenAI was able to reduce pricing across its developer platform. Highlights include 3x cheaper input token pricing for GPT-4 Turbo compared to GPT-4, and 4x cheaper input token pricing for fine-tuned GPT-3.5 compared to previous pricing.
Usage limits were also increased by 2x for GPT-4 customers. OpenAI published usage tiers to provide transparency into how limits will automatically increase with usage.
Other Announcements Other notable announcements from the event include a new Copyright Shield to defend customers against copyright claims, an experimental GPT-4 fine-tuning program, custom model development for large organizations, and upgrades like Whisper 3 for speech recognition.
Overall, OpenAI’s latest platform upgrades and reduced pricing provide developers with significantly enhanced capabilities and affordability as they build the next generation of AI applications. The powerful new GPT-4 Turbo model and multimodal integrations promise to unlock new use cases and innovations using OpenAI’s technology.
