Prove the lift · AI model performance evaluation
llama-4-maverick vs gpt-4o
+18-point lift for llama-4-maverick on ecommerce-catalog-norm
Test summary
This benchmark evaluates llama-4-maverick against gpt-4o on the ecommerce-catalog-norm v2 dataset (5 rows), scored by an LLM-as-judge (llama-3.1-8b). llama-4-maverick reached a 57% pass rate versus 39% for gpt-4o — a +18-point lift.
- Base model
- gpt-4o
- via openai · 39% pass rate
- Refined model
- llama-4-maverick
- via meta · 57% pass rate · 2.1s p95
Base pass rate
39%
Refined pass rate
57%
Refined p95
2.1s
Est. tokens saved
7.0K
+0pts
refined beats base on 5 rows
0%
Base model0%
Invoked-refined▲1 improved▼0 regressed●3 unchanged
Per-row results
1Classify a listing into the right category node100% → 100%0
2Extract material + color from a product title100% → 100%0
3Deduplicate two near-identical SKUs100% → 100%0
4Normalize a free-text size into a standard enum0% → 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