Prove the lift · AI model performance evaluation
gpt-4o vs gemini-1.5-flash
+25-point lift for gpt-4o on contract-clause-qa
Test summary
This benchmark evaluates gpt-4o against gemini-1.5-flash on the contract-clause-qa v2 dataset (5 rows), scored by an LLM-as-judge (llama-3.1-8b, run locally). gpt-4o reached a 47% pass rate versus 22% for gemini-1.5-flash — a +25-point lift.
- Base model
- gemini-1.5-flash
- via google · 22% pass rate
- Refined model
- gpt-4o
- via openai · 47% pass rate · 3.2s p95
Base pass rate
22%
Refined pass rate
47%
Refined p95
3.2s
Est. tokens saved
3.1K
+0pts
refined beats base on 5 rows
0%
Base model0%
Invoked-refined▲3 improved▼0 regressed●2 unchanged
Per-row results
1Find the liability cap and its exceptions0% → 0%0
2Determine indemnification scope and carve-outs0% → 100%+100
3Locate the termination-for-convenience terms0% → 100%+100
4Extract the auto-renewal notice window in days0% → 100%+100
5Identify the governing-law clause and jurisdiction100% → 100%0
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