FutureX: leaderboard
FutureX is ByteDance's live benchmark of future-event prediction: agents forecast outcomes of upcoming real-world events that are auto-resolved once they occur; this board reports official runs' overall score for the latest weekly round.
Metric: Overall Score (latest week, %). Source: futurex.live. Saturation forecast: Around March 2028. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 (High) | 49.19 | #31 (GPT-5.5) |
| 2 | GPT-5.4 (High) | 46.63 | #79 (GPT-5.4) |
| 3 | Qwen 3.7 Max | 44.9 | #75 |
| 4 | Seed 2.0 Pro | 42.34 | #118 |
| 5 | DeepSeek V3.2 (Thinking) | 38.48 | #205 (DeepSeek V3.2) |
| 6 | Grok 4.20 (Reasoning) | 37.17 | #175 (Grok 4.20) |
| 7 | Seed 2.0 Lite | 36.17 | #145 |
| 8 | Gemini 3.1 Pro (Preview) (High) | 35.92 | #58 (Gemini 3.1 Pro (Preview)) |
| 9 | MiniMax-M2.7 | 31.57 | #258 |
| 10 | Qwen 3.6 Max Preview | 29.47 | #85 |
| 11 | Gemini 3.5 Flash | 29.04 | #60 |
| 12 | MiniMax-M3 | 28.18 | #136 |
| 13 | DeepSeek V4 Flash | 19.56 | #137 |
| 14 | Grok 4 | 7.76 | #181 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=futurex · How It Works · Data refreshed daily, snapshot 2026-10-20.