Prove the lift · AI model performance evaluation

claude-opus-4-8 vs gemini-1.5-flash

+19-point lift for claude-opus-4-8 on ecommerce-catalog-norm

Test summary

This benchmark evaluates claude-opus-4-8 against gemini-1.5-flash on the ecommerce-catalog-norm v1 dataset (5 rows), scored by an LLM-as-judge (claude-haiku-4-5, run locally). claude-opus-4-8 reached a 43% pass rate versus 24% for gemini-1.5-flash — a +19-point lift.

Base model
gemini-1.5-flash
via google · 24% pass rate
Refined model
claude-opus-4-8
via anthropic · 43% pass rate · 1.7s p95
Base pass rate
24%
Refined pass rate
43%
Refined p95
1.7s
Est. tokens saved
9.2K
+0pts
refined beats base on 5 rows
0%
Base model
0%
Invoked-refined
0 improved0 regressed4 unchanged

Per-row results

1Extract material + color from a product title0%0%0
2Deduplicate two near-identical SKUs0%0%0
3Classify a listing into the right category node0%0%0
4Normalize a free-text size into a standard enum0%0%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