Prove the lift · AI model performance evaluation
o3-mini vs glm-5.2
+10-point lift for o3-mini on contract-clause-qa
Test summary
This benchmark evaluates o3-mini against glm-5.2 on the contract-clause-qa v1 dataset (5 rows), scored by an LLM-as-judge (llama-3.1-8b, run locally). o3-mini reached a 41% pass rate versus 31% for glm-5.2 — a +10-point lift.
- Base model
- glm-5.2
- via z-ai · 31% pass rate
- Refined model
- o3-mini
- via openai · 41% pass rate · 2.4s p95
Base pass rate
31%
Refined pass rate
41%
Refined p95
2.4s
Est. tokens saved
5.6K
+0pts
refined beats base on 5 rows
0%
Base model0%
Invoked-refined▲2 improved▼0 regressed●3 unchanged
Per-row results
1Locate the termination-for-convenience terms100% → 100%0
2Identify the governing-law clause and jurisdiction0% → 0%0
3Find the liability cap and its exceptions100% → 100%0
4Determine indemnification scope and carve-outs0% → 100%+100
5Extract the auto-renewal notice window in days0% → 100%+100
Prove the lift on your own data.
Run any two models over your tasks, judged automatically, and see the improvement — in seconds.
Try InvokedGenerated with Invoked · Browse all benchmarks