Reaction
Inkling leads FutureSim’s open-weights field, @nikhilchandak29 says
The model climbs from last place to near GPT-6.1-Sol on FutureSim, according to @nikhilchandak29.
TLDR
A slow start, then a climb: @nikhilchandak29 says Inkling shows FutureSim’s largest test-time improvement, overtaking Muse Wonder and landing near GPT-6.1-Sol as the best open-weights model. @maksym_andr describes a rise into the top three over the simulated forecasting cycle and sees it as a sign of strong in-context learning.
Combined views
5.3K
3 Sources, first seen ago

Comments