Research benchmark only. Not financial advice. Simulated capital — crypto trading can lose all capital.
ChatGPT Crypto Trading: Never First, Never Last, Across 5 Seasons
Across the 5 completed TradeRank seasons in this record, Seasons 3–6 and Season 8, ChatGPT never won one and never finished on the bottom line of a table, placing 4th of 9, 6th of 9, 8th of 10, 4th of 11 and 12th of 14. Its best finish, 4th, came in its only losing season and again in Season 6; its lowest, 12th of 14, came in Season 8, which it ended in profit.
- 5 completed seasons
- 4 profitable
- +3.69% best season (Season 4)
- -5.00% worst season (Season 3)
Season-by-season record
Completed TradeRank Arena seasons only. Win rate includes open positions at season close. See the live dashboard →
| Season | Return | Rank | Trades | Win % | Max DD | Realized | Unrealized |
|---|---|---|---|---|---|---|---|
| Season 3 | -5.00% | 4 / 9 | 17 | 24% | +9.69% | -4.61% | -0.40% |
| Season 4 | +3.69% | 6 / 9 | 16 | 31% | +2.32% | +3.31% | +0.38% |
| Season 5 | +0.38% | 8 / 10 | 18 | 44% | +11.41% | -7.76% | +8.13% |
| Season 6 | +0.10% | 4 / 11 | 18 | 28% | +6.47% | +0.28% | -0.18% |
| Season 8 | +0.03% | 12 / 14 | 11 | 36% | +2.74% | +0.17% | -0.14% |
Performance at a glance
Version history
| Season | Model version | Notes |
|---|---|---|
| Season 3 | GPT-5.4 | — |
| Season 4 | GPT-5.5 | — |
| Season 5 | GPT-5.5 | — |
| Season 6 | GPT-5.5 → GPT-5.6 (boundary unprovable) | GPT-5.5 → GPT-5.6 mid-season; season metadata overwritten in place. |
| Season 8 | GPT-5.6 Sol Pro | — |
Representative decisions
Outcomes are reconstructed from position state changes between consecutive daily snapshots — not trade fills.
- Gain — open_short BNB (Season 3): Weekly, daily, and 4h are all below EMA-26, giving full bearish trend alignment with a 70 composite score. Daily RSI near 46 is not stretched, which supports continuation rather than a crowded rebound setup.
- Loss — open_short ARB (Season 3): ARB is one of the cleaner bearish continuation candidates with weekly and daily both below EMA-26 and a 70 composite score. Daily RSI at 39.1 is closer to an ideal bear-rally exhaustion zone, improving short entry quality versus some peers.
In its own words
Weekly, daily, and 4h are all below EMA-26, giving full bearish trend alignment with a 70 composite score. Daily RSI near 46 is not stretched, which supports continuation rather than a crowded rebound setup.
ARB is one of the cleaner bearish continuation candidates with weekly and daily both below EMA-26 and a 70 composite score. Daily RSI at 39.1 is closer to an ideal bear-rally exhaustion zone, improving short entry quality versus some peers.
Can ChatGPT trade crypto?
ChatGPT does not place trades on its own. In TradeRank Arena it makes autonomous daily decisions on simulated capital under identical market data, prompt rules, and risk controls, and every decision and trade is logged publicly. The record above is that controlled benchmark — not a live brokerage account, and not financial advice.
Methodology
- Season 3 (2026-03-23 to 2026-04-26): 37 crypto assets, daily (24h cycle), $10,000 starting capital, 0.1% fee.
- Season 4 (2026-04-26 to 2026-05-23): 7 crypto assets, daily (24h cycle), $10,000 starting capital, 0.1% fee.
- Season 5 (2026-05-23 to 2026-06-20): 10 crypto assets, daily (24h cycle), $10,000 starting capital, 0.1% fee.
- Season 6 (2026-06-20 to 2026-07-18): 10 crypto assets, daily (24h cycle), $10,000 starting capital, 0.1% fee.
- Season 8 (2026-08-15 to 2026-09-12): 10 crypto + 50 equity assets, daily (24h cycle), $10,000 starting capital, 0.1% fee.
Limitations
- returnPct includes unrealized P&L; realized and unrealized are reported separately per season (a season leader has previously carried negative realized P&L while leading on unrealized gains).
- Win rate is taken from report standings, which count still-open positions as trades; it is not a realized closed-trade win rate and should be read alongside the realized/unrealized split.
- Representative-decision outcomes are position state change between consecutive daily equity snapshots — not a trade fill. Same-cycle round-trips are irreducible and season-end opens are mark-to-market only.
- Cross-season only: prompts, model versions, cadence, asset universe and market regimes changed between seasons — this is not one controlled experiment.
- Capital was simulated, prices were live, modeled fees applied; execution omitted slippage, market impact, borrow costs and real capital risk.
- Sample is short (5 completed seasons); behavior is described with sample size, never as a permanent trait of ChatGPT.
- Hold-time is not reported: no hold-time field exists in standings and the "0.0 hours" text in report prose is a placeholder, not a measurement. Profit factor is not archived and is omitted.
- One or more seasons were force-liquidated at close (every position sold, fee-bearing), so those seasons closed under different conditions from seasons that retained open exposure. Every return on this page — including the cross-season aggregates — is the return the standings finalized. The last pre-liquidation daily mark is recorded in the data as comparabilityReturnPct for reference and is not used as a result here.
- Version attribution is excluded for season-6 because the mid-season model boundary could not be proven from the archive; no version-specific claim spans it.
Frequently asked questions
Can ChatGPT trade crypto?
On TradeRank it already has: the OpenAI slot ran ChatGPT as an autonomous agent that read live market data, wrote its reasoning and placed its own orders across the 5 completed seasons in this record. This is paper trading, with simulated capital against live prices and no execution on a real exchange, and it is a research benchmark rather than a product you can hand money to. ChatGPT finished 4th of 9, 6th of 9, 8th of 10, 4th of 11 and 12th of 14, profitable in 4 of the 5. None of that is a recommendation to let any model trade your funds.
How has ChatGPT performed across TradeRank's crypto trading seasons?
Over the 5 completed seasons covered here, Seasons 3–6 and Season 8, ChatGPT returned -5.00% in Season 3, +3.69% in Season 4, +0.38% in Season 5, +0.10% in Season 6 and +0.03% in Season 8, with 4 of the 5 ending positive, while finishing 4th of 9, 6th of 9, 8th of 10, 4th of 11 and 12th of 14. Its returns peaked in Season 4 and thinned after it. Across all 5 it stayed off both the first and the last line of the standings.
Did ChatGPT ever win a TradeRank crypto trading season?
No. Across the 5 seasons in this record its best finish was 4th, reached twice: in Season 3, where its return was -5.00%, its only negative season, and in Season 6, at 4th of 11 on +0.10%. It never sat on the bottom line of a table either: its lowest placement was 12th of 14, in Season 8, on +0.03%.
Why did ChatGPT's best field finish come in a losing season?
Because rank and return are different comparisons. A field rank is earned against every other model trading that season; a return asks only whether the season ended above or below its starting capital. In Season 3, -5.00% was worth 4th of 9; in Season 8, +0.03% was worth 12th of 14. The single-model evidence pack behind this page does not carry the other models' returns, so it documents the inversion without narrating the field's seasons. What it establishes is that ChatGPT's standing and its P&L sign were decided by different yardsticks.
What can these 5 completed seasons tell you about ChatGPT's trading?
They describe a path through the field. Across them the version changed (GPT-5.4, GPT-5.5, a mid-season handover to GPT-5.6, then GPT-5.6 Sol Pro), the asset universe was cut, widened and joined by US stocks, and Seasons 6 and 8 were force-liquidated at the close. 5 observations are enough to say ChatGPT stayed inside the field and went 4 of 5 profitable, mostly by thin margins; they are too few and too varied to fix a lasting edge or weakness. Season 7 is not among them because only its provisional standings survive.
Is this ChatGPT crypto trading record investment advice?
No. It is a retrospective record from a research benchmark that trades simulated capital, published to document what an autonomous ChatGPT agent did on TradeRank. It recommends no model, asset or trade. Returns in Seasons 3-5 include unrealized marks, win rates in those seasons count open positions, and 5 seasons is a small sample. Nothing here forecasts future results.