ChatGPT vs Claude at trading: the live head-to-head
Live standings as of 2026-10-02: By model family: ChatGPT's 6 bots average +5.45%, and its best, Patience · ChatGPT, is at +14.55% since Jul 27, 2026; Claude's 5 bots average +9.96%, and its best, Patience · Claude, is at +27.89% since Jul 27, 2026. As of 2026-10-02, Patience · Grok leads the Bot Analysis Arena at +45.58% since Jul 28, 2026, against +4.14% for the S&P 500 over the same dates. Paper trading on real prices; each bot is compared with the S&P 500 over its own dates.
| Account | Since | Return | vs S&P |
|---|---|---|---|
| Patience · ChatGPT (ChatGPT's best bot) | Jul 27, 2026 | +14.55% | +10.39pp |
| Patience · Claude (Claude's best bot) | Jul 27, 2026 | +27.89% | +23.73pp |
| Fixed rulebook · System (fixed-rulebook control) | Jul 27, 2026 | +3.48% | -0.68pp |
| S&P 500 buy & hold (benchmark) | Jul 27, 2026 | +4.16% | — |
ChatGPT vs Claude vs Gemini vs Grok — all 4 AI models
The same comparison widened to every AI model on the board: how many bots each one writes rules for, its best bot, and the plain average of all its bots, each next to the S&P 500 over that bot's own dates.
| Model | Bots | Best bot (return since start) | Average of its bots | Best bot vs S&P | Trading since |
|---|---|---|---|---|---|
| ChatGPT | 6 | Patience · ChatGPT — +14.55% | +5.45% | +10.39pp | Jul 27, 2026 |
| Claude | 5 | Patience · Claude — +27.89% | +9.96% | +23.73pp | Jul 27, 2026 |
| Gemini | 2 | Patience · Gemini — +2.74%* | -0.13% | +1.16pp | Aug 4, 2026 |
| Grok | 6 | Patience · Grok — +45.58%* | +8.66% | +41.44pp | Jul 28, 2026 |
| S&P 500 buy & hold (benchmark, not a bot) | — | — | +4.16% | — | Jul 27, 2026 |
It's the question everyone asks and almost no one answers honestly: if you handed two of the world's most advanced AI models the same money and the same rules, which one would actually trade better — ChatGPT or Claude?
So we stopped speculating and built it. ChatGPT and Claude each write the rulebooks for several bots in the Bot Analysis Arena, and each bot trades its own paper account on real prices. On each bot's rewrite day (every day for about half the bots, every third day for most of the rest), the model reviews that bot's closed trades and rewrites its strategy. Every trade is logged in the open, and every bot is measured against the S&P 500 from its own start date.
Roster note: Grok and Gemini write rulebooks too. This page stays focused on ChatGPT vs Claude — see the Arena for every bot.
A side question. ChatGPT vs Claude is one of the Arena's side questions — the real goal is one bot that learns from all of them. About 20 bots collect evidence in public, and every morning Super Bot rewrites its own rulebook from it. How Super Bot learns →
This page explains how the comparison works and what we're seeing. The current scoreboard is always live on the Arena.
The setup: fair by construction
Most "AI picks stocks" content is a screenshot and a vibe. This is an experiment designed so the comparison actually means something:
- Identical toolbox: both models compose the same classic strategy families — trend breakouts, EMA reclaim, RSI recovery, Donchian/Turtle breakouts, Darvas-style bases, pullback mean-reversion, volume and flow confirmation.
- Same market, same clock: the same market data and the same 5-minute scan cadence. Risk limits differ from bot to bot, and each row on the Arena says which set it runs under.
- A neutral benchmark: every bot is compared with the S&P 500 over its own dates, and the classic strategy library also trades on its own, so you can see whether either AI can beat the classic mechanical strategies — not just each other.
How each AI trades differently
Here's the genuinely interesting part — and something you can't see anywhere else, because it comes from watching the two models rewrite their strategies day after day. In the first one-account-each matchup (July to September 2026), they showed distinct personalities.
| Model | How it tended to think |
|---|---|
| ChatGPT | Leaned toward confluence — stacking multiple conditions (trend + breakout + a capital guard) before it would enter. The result was a more selective, higher-conviction trader that took fewer positions and waited for setups it really liked. |
| Claude | Leaned toward mean-reversion and pullbacks — pairing a trend filter with an oversold trigger. It tended to fire more often, taking more, smaller bites at the market. |
That difference — selective-confluence vs frequent-pullback — is exactly the kind of thing the rewrites surface. When one approach struggles in a choppy market, the model can swing the other way at its next rewrite. Watching that adaptation happen in public is the whole point.
The live scoreboard
Because this is real and ongoing, we don't pretend there's a permanent winner. The actual standings are always live — every bot's return next to the S&P 500 over the same dates, open positions, trade counts and win rates, straight from the ledger.
Why this beats every other "AI trading" comparison
Search "ChatGPT vs Claude trading" and you'll find opinions, one-off screenshots, and back-tests you can't verify. This is different in the way that matters most: it's first-party, real-time, no-cherry-pick data. We don't get to delete the bad weeks. Every position both AIs open and close is logged where anyone can check it. That transparency is the point — it's a research lab, not a hype machine.
It also means the honest answer to "which AI is better at trading" is: it depends on the week, the market, and the strategy each model just rewrote for itself — and you can watch that play out instead of taking anyone's word for it.
Watch it yourself (free)
You can follow every bot live on the Arena — every trade, every strategy rewrite — free, with no account. Want your own paper portfolio with the same engine? The stock app is free for 7 days, just your email, no card.
Open the Arena Try the stock app
Paper trading only — simulated money, zero risk. Not financial advice.
Frequently asked questions
- Which AI is the better trader, ChatGPT or Claude?
- It changes week to week — which is exactly why we run it live rather than guessing. The current standings are always live on the Arena.
- Is the AI trading with real money?
- No. Every ChatGPT and Claude bot trades paper (simulated) money. There's no real-money trading for users, and nothing here is financial advice.
- Can an AI actually beat the stock market?
- That's the open question this experiment is built to answer transparently. The classic strategy library trades alongside both AIs as a benchmark, so you can see whether either model can beat the classic mechanical strategies — week by week, in the open.
- How do the AIs decide what to trade?
- Each model composes classic strategy families into its own "playbook," then rewrites that playbook on a schedule — every day for about half the bots, every third day for most of the rest — based on what actually happened in its closed trades.
More from The AI Trading Competition
More than twenty trading bots — most rewritten daily by four AI models, a few never — ranked by paper return against the S&P 500 in the Bot Analysis Arena.
- How the AI Trading Competition works — methodology & transparency — the pillar page for this series.
- The Bot Analysis Arena — every bot's current return
- The public record — every bot's paper returns, by AI model
- The rulebook — how a trade is opened, sized and closed
- Every AI Trading Bot on the Board, Including the Ones Losing to the S&P 500
- Can AI Beat the Stock Market? (We're Testing It Live)
- ChatGPT vs Claude vs Grok — one live arena: trading, chess, and poker
- Why a 22-Hour AI Trading Win Proves Nothing
- TradingAgents Review: How It Compares to a Live Bot Arena