OpenAI Finally Introduces NEW MODELS! (Updated GPT-4) + Leaked Updates

1. OpenAI introduces updated GPT-4 and GPT-3.5 models.

🥈85 00:19

OpenAI has released an updated GPT-4 Turbo preview model and an upgraded GPT-3.5 Turbo model with lower pricing, reduced by 50 to 0.005 cents per 1,000 tokens.

• The GPT-3.5 Turbo model will have various improvements, including higher accuracy at responding in requested formats and a fix for a bug causing a text encoding issue for non-English language functional calls.
• Customers using the pinned GPT-3.5 Turbo model will be automatically upgraded from the GPT-3.5 Turbo 0613 to GPT-3.5 Turbo 0125 two weeks after the new model launches.

2. Enhancements in GPT-4 Turbo model to reduce laziness and improve performance.

🥇92 02:29

The updated GPT-4 Turbo preview model aims to reduce cases of laziness where the model doesn't complete a task thoroughly, providing a larger 128k context window and lower prices.

• Over 70% of requests from the GPT-4 API customers have transitioned to GPT-4 Turbo since its release, taking advantage of its updated knowledge cut-off and larger context window.
• The updated GPT-4 aims to address documented issues of laziness compared to the previous version, as evidenced by a Stanford study.

3. Improved moderation model and API key management.

🥈88 05:55

OpenAI has released an updated moderation model and new ways to understand API usage and manage API keys, providing developers with more visibility and control over API keys.

• The free moderation API allows developers to identify potentially harmful text, and the usage dashboard and export function now exposes metrics on an API key level after turning on tracking.
• Developers can assign permissions to API keys from the API key page, allowing restricted access to certain endpoints and more control over usage.

4. Future features include personalization and memory feature for GPT-3.

🥈82 07:32

OpenAI is working on future features for GPT-3, including personalization and memory features, allowing the ability to mention a GPT and add it directly into a conversation.

• The ability to mention a GPT and add it directly into a conversation is a potential future feature, enhancing workflow and enabling the use of multiple GPTs in different conversations.
• These future features are likely to enhance the effectiveness and versatility of GPT-3 in various contexts.