Gemini 3.0 Prop Bets
20
Ṁ1399Dec 31
89%
new SOTA at FrontierMath Tier 4 from some gemini 3 model (deepthink counts)
65%
new SOTA at FrontierMath Tiers 1-3 from some gemini 3 model (deepthink counts)
63%
>=70% SWE-bench Verified (Bash only, per swebench.com)
60%
2M+ context window on Gemini 3.0 Pro
56%
Gemini app tops App Store (#1) within 1 month of Gemini 3 release
50%
new SOTA at MMLU
48%
DeepThink version released within 2 months of Gemini 3.0 Pro release
46%
Gemini 3.0 Pro METR time horizon of 3 hours or more
41%
Gemini 3.0 Flash released before Gemini 3.0 Pro (includes experimental versions)
25%
DeepThink version of model released alongside Gemini 3.0 Pro
17%
native image-out capabilities on launch
15%
Gemini 3.0 Ultra announced in main Gemini 3.0 announcement
This question is managed and resolved by Manifold.
Get
1,000and
3.00
1,0003.00