Prove the lift · AI model performance evaluation

llama-3.1-70b vs mistral-large

+14-point lift for llama-3.1-70b on ecommerce-catalog-norm

Test summary

This benchmark evaluates llama-3.1-70b against mistral-large on the ecommerce-catalog-norm v1 dataset (5 rows), scored by an LLM-as-judge (gpt-4o-mini, run locally). llama-3.1-70b reached a 53% pass rate versus 39% for mistral-large — a +14-point lift.

Base model
mistral-large
via mistral · 39% pass rate
Refined model
llama-3.1-70b
via ollama · 53% pass rate · 3.6s p95
Base pass rate
39%
Refined pass rate
53%
Refined p95
3.6s
Est. tokens saved
7.4K
+0pts
refined beats base on 5 rows
0%
Base model
0%
Invoked-refined
1 improved0 regressed3 unchanged

Per-row results

1Deduplicate two near-identical SKUs100%100%0
2Extract material + color from a product title100%100%0
3Classify a listing into the right category node0%100%+100
4Normalize a free-text size into a standard enum100%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 Invoked
Generated with Invoked · Browse all benchmarks