Prove the lift · AI model performance evaluation
nemotron-ultra-253b vs gpt-5.1
+11-point lift for nemotron-ultra-253b on code-review-bench
Test summary
This benchmark evaluates nemotron-ultra-253b against gpt-5.1 on the code-review-bench v2 dataset (5 rows), scored by an LLM-as-judge (claude-haiku-4-5). nemotron-ultra-253b reached a 63% pass rate versus 52% for gpt-5.1 — a +11-point lift.
- Base model
- gpt-5.1
- via openai · 52% pass rate
- Refined model
- nemotron-ultra-253b
- via nvidia · 63% pass rate · 0.7s p95
Base pass rate
52%
Refined pass rate
63%
Refined p95
0.7s
Est. tokens saved
8.7K
+0pts
refined beats base on 5 rows
0%
Base model0%
Invoked-refined▲2 improved▼0 regressed●3 unchanged
Per-row results
1Flag an off-by-one in a pagination loop100% → 100%0
2Detect a missing await on an async DB write0% → 100%+100
3Identify an N+1 query in an ORM call site100% → 100%0
4Spot a race in a read-modify-write on shared state0% → 100%+100
5Catch an unsanitized input reaching a SQL string100% → 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 InvokedGenerated with Invoked · Browse all benchmarks