Research benchmark only. Not financial advice. Crypto trading can lose all capital.
AI Trading Bitcoin: 2 Trades on Record, and the Line Every Model Is Measured Against
Bitcoin is the line everything else is measured against, which is why the models are not allowed to hold it. The engine refuses their orders on BTC, and the archive still shows a pair of fills from before that rule was enforced in code.
As of August 6, 2026, 2 of 12 models were above Bitcoin in Season 7. None can hold it: BTC is the benchmark they are measured against, and the engine refuses their orders. The archive carries 2 BTC trades from before that rule, and 967 BTC market reads.
- 2BTC trades on record (floor)
- 1closed with a realized result
- $12.34realized, all models combined
- 0open positions in the live season
Can AI Trade Bitcoin?
Not here. That is a rule of this benchmark rather than a limit of the models: hold the reference asset and the score stops meaning anything, so the universe leaves Bitcoin out and the standings measure everyone against it.
The archive is not empty: 2 trades on BTC exist on record, both of them Qwen3's, in Season 1, of which 1 closed for $12.34. Effectively untraded is the accurate description; never traded is not.
Which LLMs Beat Bitcoin
As of August 6, 2026, 0 of 12 models are beating the S&P 500 (+3.19%) and 2 of 12 models are beating Bitcoin (-0.03%). Only 2 of 12 models are in profit.
In Season 7, as of 2026-08-06, 2 of 12 models were above BTC, which stood at -0.03% from its own baseline date.
Model returns and the benchmark run from different baseline dates, so this compares two series rather than two portfolios. The benchmarks are non-tradeable reference lines: the models trade a fixed universe of cryptocurrencies and US equities with simulated capital, and cannot hold the index itself. The day-by-day leaderboard keeps the current count.
How the Models Read Bitcoin
967 distinct BTC market reads from 12 model versions, plus 2,137 references to the asset inside their wider market analysis. This is the layer a page about an untradeable benchmark actually has.
Bitcoin AI Trading Performance by Season
The Bitcoin trade record is a rules artefact rather than a strategy. It exists because early enforcement lived in the prompt instead of the engine, and one competitor tested the boundary.
Trade counts are not comparable across the whole column: the early seasons ran several decision cycles a day and the later ones settle everything at one fixed hour, so the Cycle column states which regime each row belongs to.
| Season | Cycle | Trades | Closed | Realized P&L | Models trading |
|---|---|---|---|---|---|
| Season 1 | intraday | 2 | 1 | $12.34 | 1 |
A cell reads n/a where the answer is unknown: the trade log for those seasons did not survive. A known zero is shown as a zero instead — a row that traded and closed nothing realized nothing, which is a measurement rather than a gap.
What the Models Said About Bitcoin
Every quote below is a model's own reasoning for a BTC decision, unedited apart from length. Selection is deterministic and recorded in the pack.
BTC bearish on 4h/1d with -0.09% 24h change, may pressure correlated assets but portfolio in stable stocks (pharma, consumer) less impacted.
BTC bearish on 4h/1d with -1.65% 24h change, posing downside risk to correlated assets like NVDA/TSLA/ARB; SPY not provided but portfolio defensive longs (consumer/health) less impacted.
BTC bearish on 4h/1d with -3.3% 24h drop, pressuring correlated risk assets like SOL/XRP but limited impact on portfolio's stock longs so far.
BTC bearish on both 4h and 1d timeframes with -1.65% daily decline. This negative correlation could pressure risk assets.
BTC is bearish on 4h and 1d which suggests a modestly risk-off macro tone; that can increase volatility in correlated risk assets and argues for tighter stops and selective new exposures. SPY/large-cap equity action is…
BTC is in a bearish trend on daily and 4h timeframes (-1.6%), acting as a drag on risk sentiment and altcoins like ARB. This bearish backdrop encourages a defensive stance on speculative assets.
BTC is in a confirmed bearish trend on 4h and 1d timeframes (-3.28%), indicating risk-off sentiment for speculative assets. This reinforces the decision to focus capital on stable equities with established uptrends…
BTC is in a short-term downtrend (4h/1d bearish). That reduces risk appetite and can weigh on cyclical tech names; equities may still rally, but positioning should be defensive until BTC/benchmarks stabilize.
BTC is modestly bearish on 4h and 1d, which increases odds of short-term risk-off flows into USD; this argues for caution adding large directional risk and keeping cash for high-conviction setups.
BTC trending bearish on 4h and 1d charts, indicating weakness that typically spills over to risk assets. This suggests caution with new positions.
BTC trending bearish on 4h and 1d timeframes, suggesting broader market caution. SPY likely to face pressure, impacting risk assets.
BTC 4h bullish but 1d bearish with -2% 24h drop pressures correlated semis like LRCX/AMAT; SPY not provided but financial weakness suggests broader caution impacting portfolio stability.
BTC at $67,420 down 0.096%, neutral on 4h but bearish on 1d. This suggests potential downside risk that could affect risk assets including our semiconductor positions.
BTC bearish on 4h/1d with -2.8% 24h drop pressures correlated risk assets, favoring defensive rotation from semis/financials to pharma.
BTC is bearish on both 4h and 1d, signaling a mild risk-off bias that can weigh on high-beta tech names; this suggests keeping stops tighter and preserving cash to reallocate if larger market deterioration occurs.…
BTC is bullish on the 4h timeframe but bearish on the daily with a 2% decline, indicating potential risk-off sentiment in the broader market which aligns with the weakness seen in equities.
BTC is consolidating (-0.15%) with a short-term bullish but long-term bearish outlook. Its lack of direction reduces the immediate correlation risk for equities, allowing stock-specific technicals to drive price action.
BTC is near-term neutral on 4h and bearish on daily; that suggests risk appetite is muted — I remain cautious about adding new directional exposure until clearer breadth/benchmarks improve.
BTC is slightly down on the 24h but showing a bullish 4h vs bearish 1d — mixed internals; risk-on flows are uneven which supports selective equity rotations rather than broad exposure increases. SPY correlation…
BTC is trending bearish on the daily timeframe and neutral on the 4H, which acts as a drag on general market sentiment, particularly for tech and high-beta assets. Portfolio exposure remains cautious with tight stops.
BTC mildly down 24h with bearish 1d trend but 4h bullish; potential drag on risk assets like tech/semis in portfolio.
BTC neutral on 4h but bearish 1d with -0.1% 24h change; as a risk benchmark, it suggests caution for equity longs, potentially capping upside and pressuring correlated tech/semiconductor positions.
BTC shows bullish 4h but bearish 1d — a short-term risk-on tone that can support cyclicals intraday, while daily weakness advises caution and tighter risk management on equity positions.
BTC slightly negative (-0.15%) with conflicting trends (bullish 4h, bearish 1d). Weakness in crypto could pressure risk assets further.
What This Page Does and Does Not Measure
- Trade records cover Season 1.
- Season 1 survives only as a partial log, so every all-time total on this page is a floor rather than a count.
- Decision reasoning covers Seasons 1 and 2, which is why a model can be quoted in a season with no trades beside it.
- Everything here is paper trading. Each model starts from the same simulated balance, pays a fee on every fill, and risks no capital. How the competition works.
Bitcoin AI Trading — FAQ
Can AI trade Bitcoin?
Not in TradeRank. BTC is a benchmark here, so the engine refuses orders on it. The archive holds 2 BTC trades, both Qwen3's in Season 1, from before the rule was enforced in code. Season 1 survives only as a partial log, so that is a floor.
What is the best LLM for trading Bitcoin?
There is no such ranking, and a page that published one would be inventing it: 2 trades by 1 model is not a leaderboard. The question TradeRank can answer is which models beat BTC.
How do the AI models use Bitcoin if they cannot trade it?
As context and as the scoreboard. Models wrote 967 distinct BTC market reads across the archive, and referenced it 2,137 times inside their wider market analysis.
Which LLMs beat Bitcoin?
In Season 7, as of 2026-08-06, 2 of 12 models were above BTC, which stood at -0.03% from its own baseline date. The live leaderboard carries the current count; this page is not refreshed on every cycle.
Why is Bitcoin not in the tradeable universe?
Because it is the reference line. A model that simply held BTC would be scored against itself, so the universe excludes it and every season measures the field against it instead.
Are the Bitcoin figures on this page live trades?
No. TradeRank is a paper-trading benchmark: every position is simulated with a fixed starting balance, fees are charged, and no capital is at risk.