Hermes 4 405B (Thinking) — benchmark results

Provider: Nous Research. Released 2025-08-26. Access: Open.

Unified ELO 1639 ± 32, rank #392 of 1756 rated models, from 36 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA Omniscience - Humanities & Social Sciences32.8Accuracy (%)82.9
AA Omniscience - Health31Accuracy (%)82.5
AA MMLU-Pro82.9Accuracy (%)82.3
AA Omniscience - Business27Accuracy (%)80.8
AA Omniscience - Law24.3Accuracy (%)80.8
AA-Omniscience Accuracy30.17Accuracy (%)79.8
AA Omniscience - Science, Engineering & Mathematics35.1Accuracy (%)77.1
AA LiveCodeBench68.57Pass@1 (%)76.9
AA Omniscience - Software Engineering (SWE) - Julia28Accuracy (%)75.6
AA Omniscience - Software Engineering (SWE) - R26Accuracy (%)74.6
AA Omniscience - Software Engineering (SWE) - HTML44Accuracy (%)72.1
AA Omniscience - Software Engineering (SWE) - Swift48Accuracy (%)70

Interactive version: aibenchmarks.dev/model?slug=hermes-4-405b-thinking · How the rankings work · Data refreshed daily, snapshot 2026-07-17.