OpenAI has privately announced a new developer product called Foundry, which enables customers to run OpenAI model inference at scale w/ dedicated capacity.
It also reveals that DV (Davinci; likely GPT-4) will have up to 32k max context length in the public version. 🔥
Conversation
Replying to
(GPT-3.5 Turbo appears to be referring to the ChatGPT Turbo model)
Expanded product brief and full source: docs.google.com/document/d/1Eg
read image description
ALT
14
64
217
Covered on TechCrunch 🔥
This looks to be part of a larger strategic move towards enterprise with OpenAI announcing large deals with Coca-Cola, Bain, and Spotify.
14
Replying to
See the full source doc below. Sent to me by a friend but I can’t claim to have insider confirmation.
Quote Tweet
(GPT-3.5 Turbo appears to be referring to the ChatGPT Turbo model)
Expanded product brief and full source: docs.google.com/document/d/1Eg
Show this thread
read image description
ALT
3
2
22
Show replies
Show replies
Replying to
The number of tokens that the model can "remember"
Also known as history length.
Current text-davinci-003 context length is 4096.
Larger context = more memory, recall, and smarter bots
2
2
56
Show replies
Replying to
It's interesting to know how many requests a single instance may handle. Did the document contain this information?
2
3
Replying to
No; it didn't unfortunately. probably depends on the context size of requests quite a bit