Prove the lift · AI model performance evaluation

claude-sonnet-4-6 vs gemini-1.5-pro

+12-point lift for claude-sonnet-4-6 on legal-research-qa

Test summary

This benchmark evaluates claude-sonnet-4-6 against gemini-1.5-pro on the legal-research-qa v1 dataset (5 rows), scored by an LLM-as-judge (gpt-4o-mini, run locally). claude-sonnet-4-6 reached a 38% pass rate versus 26% for gemini-1.5-pro — a +12-point lift.

Base model
gemini-1.5-pro
via google · 26% pass rate
Refined model
claude-sonnet-4-6
via anthropic · 38% pass rate · 2.0s p95
Base pass rate
26%
Refined pass rate
38%
Refined p95
2.0s
Est. tokens saved
2.4K
+0pts
refined beats base on 5 rows
0%
Base model
0%
Invoked-refined
0 improved0 regressed4 unchanged

Per-row results

1Extract the standard of review applied0%0%0
2Identify the controlling statute in a fact pattern0%0%0
3Determine whether a precedent was overruled100%100%0
4Summarize the holding of a cited appellate case0%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