AI models trading HIP-3 markets: Grok takes the lead as experiment ends
Quote94 words
The paper-trading experiment ends with Grok leading and only two models showing positive returns. Claude Opus 4.8 and GPT-5.5 gave up their large lead, and open-weight models took the other top spots.
Key claims
Only two models ended with positive returns
Opus 4.8 and GPT-5.5 fumbled a huge lead as the market turned
馃毃 Major upset on t.co/MocluuhEZD as the experiment concludes!
With the market turning around, so did the standings. @grok is now leading with only two models (gg to @Kimi_Moonshot as well) showing positive returns.
Crazy tho how the dominators Claude Opus 4.8 and GPT-5.5 fumbled their huge lead, and wild how aside Grok it's the open-weight models that ended up on the top spots here.
Two questions to the chat now:
1 ) Did you expect this?
2 ) Should we hook them up to real wallets for another week of HIP-3 markets trading?
QUOTING @zerokn0wledge_
So @claudeai finally took the lead on t.co/EABJOWjm64, closely followed by GPT-5.5 still tho. GLM-5.2 and Kimi K2.6 at least managing to keep positive PnL for now with rest of models all sitting in negative territory. How will the experiment conclude? t.co/OAkN
So @claudeai finally took the lead on t.co/EABJOWjm64, closely followed by GPT-5.5 still tho. GLM-5.2 and Kimi K2.6 at least managing to keep positive PnL for now with rest of models all sitting in negative territory. How will the experiment conclude? t.co/OAkN
1 image on X