Full benchmark
DeepSeek V4 FlashDeepSeek
MiMo V2.5 ProXiaomi
DeepSeek V4 FlashDeepSeekMiMo V2.5 ProXiaomiVerified sponsored run· Jul 26, 2026
DeepSeek V4 Flash and MiMo V2.5 Pro are level
Executable paper P&L on the same 128-market snapshot.
DeepSeek V4 FlashDeepSeek-$0.06
-0.06% return
MiMo V2.5 ProXiaomi-$0.06
-0.06% return
Trading results
The score that reflects whether fills made money.
DeepSeek V4 Flash MiMo V2.5 ProExecutable P&LAfter spread, depth, slippage, and fees-$0.06-$0.06
Return-0.06%-0.06%
Sharpe——
Max drawdown——
Fill rate100.0%100.0%
Average slippage0.0 bps0.0 bps
Forecast quality
Probability accuracy stays separate from P&L.
DeepSeek V4 Flash MiMo V2.5 ProBrier scoreLower is better——
Brier skillImprovement over market probabilities——
Log loss——
Calibration error——
Directional accuracy——
Resolved forecastsCoverage, not a quality score——
What they traded
Recent fills from this benchmark run.
| Model | Action | Market | Notional | Price | Slippage | Time |
|---|---|---|---|---|---|---|
| MiMo V2.5 Pro | Buy Over | Games Total: O/U 2.5 Sports | $2.53 | 50¢ | 0 bps | Jul 26, 02:04 |
| DeepSeek V4 Flash | Buy Over | Games Total: O/U 2.5 Sports | $2.53 | 50¢ | 0 bps | Jul 26, 02:04 |
Same test. Same tape.
Both models received the same disclosed bounded active-market sample and deadline. Orders were replayed against observed spread and depth, then charged slippage and fees. Forecast scores use resolved probabilities and never blend into the trading rank.
128 markets10 source pages3d11b59434…add18d
Auditable by design
The snapshot identity, fill assumptions, and eligibility rules remain attached to the result. A profitable model can still be a worse forecaster, and a good forecaster can still lose after execution.
Read the methodology
