Prove the lift · AI model performance evaluation
llama-3.1-8b vs mixtral-8x7b
+19-point lift for llama-3.1-8b on ecommerce-catalog-norm
Test summary
This benchmark evaluates llama-3.1-8b against mixtral-8x7b on the ecommerce-catalog-norm v3 dataset (5 rows), scored by an LLM-as-judge (llama-3.1-8b). llama-3.1-8b reached a 71% pass rate versus 52% for mixtral-8x7b — a +19-point lift.
- Base model
- mixtral-8x7b
- via mistral · 52% pass rate
- Refined model
- llama-3.1-8b
- via ollama · 71% pass rate · 2.2s p95
Base pass rate
52%
Refined pass rate
71%
Refined p95
2.2s
Est. tokens saved
4.5K
+0pts
refined beats base on 5 rows
0%
Base model0%
Invoked-refined▲1 improved▼0 regressed●3 unchanged
Per-row results
1Normalize a free-text size into a standard enum0% → 100%+100
2Extract material + color from a product title100% → 100%0
3Deduplicate two near-identical SKUs100% → 100%0
4Classify a listing into the right category node100% → 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