Verifiable-RL on VCBench
Verifiable-RL scores F0.5 29.4 on VCBench, the venture capital benchmark from the University of Oxford and Vela Research, with 29.1% precision and 30.6% recall on the 4,500-founder private test set. That is rank 9 of 32, 2.7× the F0.5 of tier-1 VCs and 3.4× Y Combinator, at $0 per 1,000 founders scored.
- Rank
- #9 of 32
- F0.5
- 29.4
- Precision
- 29.1%
- Recall
- 30.6%
- Cost / 1k founders
- $0
Scored on the 4,500-founder private test set, mean over three folds. That is 2.7× the F0.5 of tier-1 VCs. See the full leaderboard.
Compared with the reference rows
| Entry | Precision | Recall | F0.5 | Cost / 1k |
|---|---|---|---|---|
| Verifiable-RL | 29.1% | 30.6% | 29.4 | $0 |
| Think-Reason-Learn Ensemble | 40.6% | 30.1% | 37.9 | $0.87 |
| Tier-1 VCs | 23.0% | 5.2% | 10.7 | n/a |
| Y Combinator | 14.0% | 6.9% | 8.6 | n/a |
| Random Classifier | 9.0% | 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
No LLM runs at scoring time. A trained model reads structured features from the anonymized profile (education, career history, prior companies) and scores the founder directly, so the API cost at scoring is $0. Training cost is excluded, the same rule applied to every entry.
Read more
- VCBench Leaderboard Update: Accuracy vs Cost, Plus GPT-6 and Claude Opus 5.5: VCBench now plots F0.5 against the cost of scoring 1,000 founders, with GPT-6, Claude Opus 5.5, new submissions and the Think-Reason-Learn Ensemble.
Frequently asked questions
- What does Verifiable-RL score on VCBench?
- Verifiable-RL scores F0.5 29.4 on VCBench, with 29.1% precision and 30.6% recall, rank 9 of 32 as of 2026-10-07.
- Does Verifiable-RL beat human investors at predicting founder success?
- Yes. Its F0.5 is 2.7 times that of tier-1 VCs (10.7) and 3.4 times Y Combinator (8.6), after both are normalized to the dataset's 9% base rate.
- How much does Verifiable-RL cost per 1,000 founders?
- $0 at scoring time: no LLM runs when Verifiable-RL scores a founder. Training cost is excluded for every entry.