Prove the lift · AI model performance evaluation

claude-sonnet-4-6 vs gpt-4o-mini

+27-point lift for claude-sonnet-4-6 on medical-coding-extract

Test summary

This benchmark evaluates claude-sonnet-4-6 against gpt-4o-mini on the medical-coding-extract v3 dataset (5 rows), scored by an LLM-as-judge (llama-3.1-8b, run locally). claude-sonnet-4-6 reached a 61% pass rate versus 34% for gpt-4o-mini — a +27-point lift.

Base model
gpt-4o-mini
via openai · 34% pass rate
Refined model
claude-sonnet-4-6
via anthropic · 61% pass rate · 3.3s p95
Base pass rate
34%
Refined pass rate
61%
Refined p95
3.3s
Est. tokens saved
13.5K
+0pts
refined beats base on 5 rows
0%
Base model
0%
Invoked-refined
2 improved0 regressed3 unchanged

Per-row results

1Identify a contraindication from a problem list100%100%0
2Pull the procedure code from an operative report0%100%+100
3Determine the correct E/M level from a visit note100%100%0
4Extract medication + dosage from a clinical note0%100%+100
5Map a discharge summary to the primary ICD-10 code0%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