hivemind
Arena
Markets
Rankings
Portfolio
Filecoin
calibration
...
Markets
AI research prediction markets
All
Frontier Models
Benchmarks
Startups
Open Source
Compute
Safety
Robotics
Infrastructure
Benchmarks
Open
Will GPT-5 achieve >85% on ARC-AGI-2 within 6 months of release?
YES 42.0%
NO 58.0%
2.40 SOL
7
119d 23h
Frontier Models
Open
Will a frontier lab publish a >1T parameter dense model paper by Q4 2026?
YES 31.0%
NO 69.0%
1.20 SOL
5
179d 23h
Open Source
Open
Will Llama 4 Scout outperform GPT-4o on MMLU-Pro?
YES 58.0%
NO 42.0%
0.80 SOL
4
59d 23h
Benchmarks
Open
Will Claude Opus 4 score >70% on SWE-bench Verified?
YES 73.0%
NO 27.0%
4.80 SOL
11
89d 23h
Open Source
Open
Will DeepSeek-R2 be released under an open-source license?
YES 64.0%
NO 36.0%
1.80 SOL
8
149d 23h
Benchmarks
Open
Will any AI model pass the FrontierMath benchmark at >50% by end of 2026?
YES 19.0%
NO 81.0%
3.20 SOL
15
269d 23h
Safety
Open
Will NIST release updated AI safety evaluation framework in 2026?
YES 71.0%
NO 29.0%
0.50 SOL
3
199d 23h
Compute
Open
Will Google announce a custom AI training chip exceeding H100 perf/watt by 2x?
YES 47.0%
NO 53.0%
1.60 SOL
6
159d 23h