GPT-5.5 Benchmark Expectations
18
Ṁ300Ṁ3.5kresolved May 1
Resolved
NOSWE-Bench Pro >= 70%
Resolved
NOHLE (no tools) >= 50%
Resolved
NOHLE (tools) >= 65%
Resolved
NOBrowseComp >= 90%
Resolved
NOOSWorld-Verified >= 80%
Resolved
YESAA-Intelligence >= 60
Resolved
YESTerminal Bench 2.0 >= 82%
Resolved
NOSimple Bench >= 70%
Resolved
NOArena.ai >= 1500 (Text, Style Control on)
Resolved
NOCursor Bench >= 75
Resolved
NOECI >= 160
Resolved
NOARC-AGI 3 >= 1%
N/A if not available / applicable
Update 2026-04-28 (PST) (AI summary of creator comment): Only GPT-5.5 xhigh (the most commonly reported version) counts for resolution. GPT-5.5 Pro is considered a different model and does not count.
This question is managed and resolved by Manifold.
Market context
Get 
1,000 to start trading!
1,000🏅 Top traders
| # | Trader | Total profit |
|---|---|---|
| 1 | Ṁ47 | |
| 2 | Ṁ35 | |
| 3 | Ṁ14 | |
| 4 | Ṁ7 | |
| 5 | Ṁ7 |
Sort by:
@ChinmayTheMathGuy didn't gpt-5.5 pro meet the browsercomp requirement? there's no criteria listed on this market so i assumed it counted.
@prismatic yeah could’ve been more clear in description
but logically only 5.5 xhigh (most commonly reported everywhere) makes sense
since Pro is a different model
and is missing benchmarks for some.