Prove the lift · AI model performance evaluation

llama-4-maverick vs claude-sonnet-4-6

+17-point lift for llama-4-maverick on financial-doc-extract

Test summary

This benchmark evaluates llama-4-maverick against claude-sonnet-4-6 on the financial-doc-extract v3 dataset (5 rows), scored by an LLM-as-judge (claude-haiku-4-5). llama-4-maverick reached a 68% pass rate versus 51% for claude-sonnet-4-6 — a +17-point lift.

Base model
claude-sonnet-4-6
via anthropic · 51% pass rate
Refined model
llama-4-maverick
via meta · 68% pass rate · 2.6s p95
Base pass rate
51%
Refined pass rate
68%
Refined p95
2.6s
Est. tokens saved
11.3K
+0pts
refined beats base on 5 rows
0%
Base model
0%
Invoked-refined
▲1 improved▼1 regressed●3 unchanged

Per-row results

1Identify a going-concern note in the footnotes100% → 0%-100
2Pull the diluted EPS from an income statement100% → 100%0
3Extract net revenue and YoY change from an earnings PDF0% → 100%+100
4Determine the reporting period from a 10-Q header100% → 100%0
5Extract total debt from a balance sheet100% → 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