Prove the lift · AI model performance evaluation

gemini-1.5-pro vs mistral-large

+7-point lift for gemini-1.5-pro on legal-research-qa

Test summary

This benchmark evaluates gemini-1.5-pro against mistral-large on the legal-research-qa v3 dataset (5 rows), scored by an LLM-as-judge (llama-3.1-8b). gemini-1.5-pro reached a 61% pass rate versus 54% for mistral-large — a +7-point lift.

Base model
mistral-large
via mistral · 54% pass rate
Refined model
gemini-1.5-pro
via google · 61% pass rate · 2.0s p95
Base pass rate
54%
Refined pass rate
61%
Refined p95
2.0s
Est. tokens saved
2.5K
+0pts
refined beats base on 5 rows
0%
Base model
0%
Invoked-refined
1 improved0 regressed3 unchanged

Per-row results

1Determine whether a precedent was overruled100%100%0
2Identify the controlling statute in a fact pattern0%0%0
3Extract the standard of review applied0%100%+100
4Summarize the holding of a cited appellate case100%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