Nova 2.0 Pro Preview (Medium): benchmark results

Nova 2.0 Pro Preview evaluated at the medium reasoning-effort setting. Provider: Amazon. Released 2025-08-01. Access: API.

Unified ELO 1593 ± 1, rank #465 of 1761 rated models, from 22 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA IFBench79.05Accuracy (%)97.3
AA TAU-2 Bench92.69Accuracy (%)88.3
AA GPQA Diamond78.48Accuracy (%)64.5
AA Terminal-Bench Hard24.24Accuracy (%)64.5
BenchmarkList ECI121.19Capability Index (ECI)64
AA Omniscience - Humanities & Social Sciences23Accuracy (%)59.6
AA Omniscience - Law14.9Accuracy (%)59.6
AA Long Context Reasoning64Accuracy (%)59.1
Artificial Analysis Intelligence Index15.51Intelligence Index58.9
AA Omniscience - Science, Engineering & Mathematics29.7Accuracy (%)57.6
AA-Omniscience Accuracy21.97Accuracy (%)54.6
AA Omniscience - Health21Accuracy (%)52.6

Interactive version: theaggregate.ai/model?slug=nova-2-0-pro-preview-medium · How It Works · Data refreshed daily, snapshot 2026-09-05.