Development

A model beats the S&P with real money

development2026-01-26

Claim

EconBench found GPT-5 most economically rational and Sonnet 4.5 most prosocial, while Rallies AI's live test had Sonnet 4.5 returning 8.7% against the S&P's 1.9% since November.

Reported metric: 8.7% vs 1.9%

Reporting date: 2026-01-26 (not necessarily the event date)

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

Welcome to January 26, 2026 (theinnermostloop.substack.com)

Knowledge graphopen in browser →

About

EconBench

Reported in

Welcome to January 26, 2026

Actor

Rallies AI

Evidence for

Supersedes

A benchmark gives models $10,000 to bet

Referenced by

Cite or correct this record

Citation formatCitation text

A model beats the S&P with real money
Innermost Loop KB — secondary record
Reporting date: 2026-01-26
KB permalink: https://nicholsn.github.io/innermost-loop-kb/developments/2026-01-26-econbench-and-live-trading
Claim: EconBench found GPT-5 most economically rational and Sonnet 4.5 most prosocial, while Rallies AI's live test had Sonnet 4.5 returning 8.7% against the S&P's 1.9% since November.
Attributed sources:
- Welcome to January 26, 2026 — https://theinnermostloop.substack.com/p/welcome-to-january-26-2026
This citation does not imply independent verification of linked-source claims.

Suggest a correction — opens a GitHub draft for you to review and submit.