Capability gains

OpenAI deploys fill-in-the-middle code model code-davinci-002

OpenAI described a training method that added fill-in-the-middle capability without harming left-to-right performance and identified code-davinci-002 as its deployed infilling model.

CURRENT ASSESSMENT · REVISION 1
TOWARD DOOM48confidence 96/100

Why it moved the index

The result added a useful code-editing capability to a model already released through the API, directly improving deployed software automation. The work concerned targeted code infilling rather than broad autonomous agency.

0 comments · 0 votesOpen discussion

Public discussion is readable by everyone. Sign in to comment, reply, or vote.

No comments yet. Start the discussion.

AUDIT TRAIL

Assessment history

  1. R1
    Toward 48 · confidence 96

    New July 2022 exact deployed code-model result absent from the durable corpus.

    12 Aug 2026