OpenAI Enhances GPT-4 with Turbo Update to Combat Performance Decline

Recent external research on OpenAI’s GPT-3.5 and GPT-4 models revealed a concerning trend: over time, their performance and behavior have deteriorated significantly. Addressing these findings, OpenAI has introduced the GPT-4 Turbo update, aimed at revitalizing the model’s efficiency, particularly in code generation tasks. This update specifically targets reducing instances of "laziness" where the AI fails to complete assigned tasks.

While OpenAI’s announcement brought attention to the improvements, details on the underlying issues and the specific modifications made remain undisclosed. Furthermore, the launch of GPT-4 Turbo with vision (GPT-4 V) is anticipated to reach general availability in the near future, promising even broader capabilities.

In addition to the GPT-4 enhancements, OpenAI has rolled out an updated version of GPT-3.5 Turbo (gpt-3.5-turbo-0125), featuring various refinements, including a correction for a text encoding bug affecting non-English language function calls.

OpenAI also declared another price reduction for GPT-3.5 Turbo, marking the third cut within a year. The new pricing model significantly lowers costs, with input prices halving and output prices decreasing by 25%.

Since the debut of ChatGPT, OpenAI's generative AI models have set the benchmark for AI performance, influencing the development strategies of competitors worldwide. For instance, iFlytek, a Chinese AI company, has announced plans to upgrade its Spark Model to rival GPT-4 Turbo’s capabilities by the year's midpoint.

iFlytek's Spark V3.5 is reported to closely match, and in some respects, exceed GPT-4 Turbo’s performance in language understanding and mathematical tasks.

What aspect of the GPT-4 Turbo update do you find most promising?

• Enhanced Task Completion
• Introduction of GPT-4 Turbo with Vision
• Improved Non-English Language Support
• Significant Price Reduction

See All Options