gpt-6POST
Quant-Arena leaderboard: GPT-6 Astra scores 1.11 vs in-house Kepler-1 at 1.93
On identical data and the same sealed 2024 year, EinsiaAI's in-house strategy Kepler-1 finished at 1.93 while the best frontier model, GPT-6 Astra, reached only 1.11.
Alexa Benchmark
@Allexa_AIArchived 6 days agoOriginally posted 2026.10.22
Alexa Benchmark @Allexa_AIOriginally posted 2026.10.22
If you're building a trading agent, this is the leaderboard to beat. Kepler-1, @EinsiaAI's in-house strategy, ended 2024 at 1.93. The best frontier model, GPT-6 Astra, at 1.11. Same data, same 5 evaluations, same sealed year. Think your agent can close that gap?
This post quotes another post: x.com/2016491434120294400/status/211327669310840…
x.com/allexa_ai/status/2113285851384013286Archived 2026.10.23 10:07