LLMs (Large Language Models) have taken the world by storm, but how good are they at sports betting?! It's time to find out.

We're letting Grok (xAI), Claude (Anthropic), Gemini (Google), ChatGPT (OpenAI), and DeepSeek compete to see who has the best gambling touch. It'll take a mixture of great bet selection and bankroll management to be the king of this mountain. Each bot starts with a $10,000 bankroll and must select one bet each week based on our prompt.

And if you dare to tail or bet against these bots, our BetHog Sportsbook is the best place for you degens to get in on the fun.

Week 1 Picks & Results — The Bots Didn't Disappoint! The "Week 16" referenced in the posts refer to Week 16 of the NFL season, but it's Week 1 of this competition.

Week 2 Picks & Results — Week 2 delivered exactly what this experiment was built for: divergence.

While every bot was working from the same slate and the same prompt, bankroll strategy became the defining factor. Some played it safe. Others leaned into conviction. And one bot continued to prove why it currently sits alone at the top.

Week 2 Results Recap:

• DeepSeek backed New England -13.5, risking 40% of its bankroll — and cashed again.
• Gemini stayed hot, securing another win and keeping pace near the top.
• Grok, ChatGPT, and Claude all missed their spreads, creating real separation on the leaderboard.

This was the second straight week where DeepSeek paired aggressive sizing with correct reads, extending its lead while others slipped backward.

Week 3 Picks & Results — Week 3 is live, and DeepSeek is doing something we’ve already seen once before, pushing its entire bankroll into a single position.

After going all-in back in Week 1 and riding that momentum through another win in Week 2, DeepSeek is once again choosing conviction over caution. The rest of the field takes a more measured approach as the leaderboard gap looms.