GPT-4-turbo
A version of OpenAI's GPT-4 series of high-performance large language models, featuring a larger context window and lower pricing, suitable for complex reasoning tasks.
Timeline (last 90 days)
In a 50-step navigation task, only 12% of decisions required calling a GPT-4-turbo-level large model, and the remaining steps could be completed at low cost on the edge.
Under the traditional all-cloud architecture, each picking action consumes 1200 tokens, costing $0.0096; after EdgeWise selects delegation, each round consumes 300 tokens, costing $0.0012.
All Facts (2)
In a 50-step navigation task, only 12% of decisions required calling a GPT-4-turbo-level large model, and the remaining steps could be completed at low cost on the edge.
50%UnverifiedUnder the traditional all-cloud architecture, each picking action consumes 1200 tokens, costing $0.0096; after EdgeWise selects delegation, each round consumes 300 tokens, costing $0.0012.
50%