← All models

o3 on VCBench

o3 scores F0.5 21.5 on VCBench, the venture capital benchmark from the University of Oxford and Vela Research, with 43.2% precision and 7.4% recall on the 4,500-founder private test set. That is rank 24 of 34, 2.0× the F0.5 of tier-1 VCs and 2.5× Y Combinator, at $9.25 per 1,000 founders scored.

o3 on the VCBench leaderboard today
Rank
#24 of 34
F0.5
21.5
Precision
43.2%
Recall
7.4%
Cost / 1k founders
$9.25

Scored on the 4,500-founder private test set, mean over three folds. That is 2.0× the F0.5 of tier-1 VCs. See the full leaderboard.

Compared with the reference rows

EntryPrecisionRecallF0.5Cost / 1k
o343.2%7.4%21.5$9.25
Think-Reason-Learn Ensemble40.6%30.1%37.9$0.87
Tier-1 VCs23.0%5.2%10.7n/a
Y Combinator14.0%6.9%8.6n/a
Random Classifier9.0%9.0%9.0$0

Human rows are normalized to the dataset's 9% success rate, so every entry is compared on the same base rate. Costs use list prices on 2026-09-24.

How it was scored

An LLM reads each founder profile at scoring time and returns a prediction. o3 used about 284 input and 1,085 output tokens per founder on o3, estimated from the prompt and the saved responses; reasoning models are assumed to think about 1,000 tokens per founder. Cost is tokens times the provider's list price on 2026-09-24 ($2 input and $8 output per million tokens), so it recomputes when prices change. Build costs such as question or policy generation are excluded.

Read more

Frequently asked questions

What does o3 score on VCBench?
o3 scores F0.5 21.5 on VCBench, with 43.2% precision and 7.4% recall, rank 24 of 34 as of 2026-10-26.
Does o3 beat human investors at predicting founder success?
Yes. Its F0.5 is 2.0 times that of tier-1 VCs (10.7) and 2.5 times Y Combinator (8.6), after both are normalized to the dataset's 9% base rate.
How much does o3 cost per 1,000 founders?
About $9.25 per 1,000 founders at list prices on 2026-09-24. An LLM reads each founder profile at scoring time and returns a prediction. o3 used about 284 input and 1,085 output tokens per founder on o3, estimated from the prompt and the saved responses; reasoning models are assumed to think about 1,000 tokens per founder. Cost is tokens times the provider's list price on 2026-09-24 ($2 input and $8 output per million tokens), so it recomputes when prices change. Build costs such as question or policy generation are excluded.