Prove the lift · AI model performance evaluation
claude-opus-4-6 vs claude-opus-4-8
+18-point lift for claude-opus-4-6 on insurance-extraction-eval
Test summary
This benchmark evaluates claude-opus-4-6 against claude-opus-4-8 on the insurance-extraction-eval v3 dataset (7 rows), scored by an LLM-as-judge (gpt-4o-mini). claude-opus-4-6 reached a 65% pass rate versus 47% for claude-opus-4-8 — a +18-point lift.
- Base model
- claude-opus-4-8
- via anthropic · 47% pass rate
- Refined model
- claude-opus-4-6
- via anthropic · 65% pass rate · 3.3s p95
Base pass rate
47%
Refined pass rate
65%
Refined p95
3.3s
Est. tokens saved
8.6K
+0pts
refined beats base on 7 rows
0%
Base model0%
Invoked-refined▲1 improved▼1 regressed●5 unchanged
Per-row results
1Suit-limitation clause: is a late-filed claim contractually barred?100% → 100%0
2Business-interruption period-of-restoration calculation100% → 100%0
3Coinsurance 80% clause recovery on an under-insured building100% → 0%-100
4Aggregate deductible across a loss sequence under a per-occurrence + annual cap0% → 100%+100
5Subrogation waiver interaction with additional-insured status100% → 100%0
6Named-peril coverage determination under a basic HO form100% → 100%0
7Pro-rata distribution across layered carriers on a covered loss0% → 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 InvokedGenerated with Invoked · Browse all benchmarks