Models / Nous
Hermes 4 405B
by Nous ·
nousresearch/hermes-4-405bHermes 4 is a large-scale reasoning model built on Meta-Llama-3.1-405B and released by Nous Research. It introduces a hybrid reasoning mode, where the model can choose to deliberate internally with...
reasoningreleased 2025-08-26
Legal score · raw
-
not yet benchmarked
Context
131K
max output 118K
Input
$1
per 1M tokens
Output
$3
per 1M tokens
Suite cost
-
run the suite to see
Benchmark results
| Task | Category | Raw | Juiced | Correct | Latency | Cost | Ran |
|---|---|---|---|---|---|---|---|
| Hearsay Identification | Evidence | - | - | - | - | - | - |
| Bluebook Citation Format | Research & Writing | - | - | - | - | - | - |
| Federal Civil Procedure | Procedure | - | - | - | - | - | - |
| Limitations Arithmetic | Procedure | - | - | - | - | - | - |
| Contract Clause Classification | Contracts | - | - | - | - | - | - |
| Citation Hallucination Resistance | Reliability | - | - | - | - | - | - |
Measured by DocketBuster
These numbers come from DocketBuster's own legal battery, not from DocketRouter's suite. Latest run per metric, with n and a 95% Wilson interval where the source reports one. See docketbuster.com/benchmarks.
| Metric | Value | n | Interval | Measured |
|---|---|---|---|---|
| Statute pinpoint, exact section (no retrieval) | 7.7% (23/300) | 300 | 95% CI 5.2% to 11.2% | 2026-08-22 |
| Statute pinpoint, exact section (with DocketBuster retrieval) | 75.6% (232/307) | 307 | 95% CI 70.5% to 80.0% | 2026-08-24 |
| Say-nothing rate (declines to bluff when the answer is not in the record) | 99.6% | 244 | 95% CI 97.7% to 99.9% | 2026-08-22 |
| Abstained on statute pinpoint | 16.3% (50/307) | 307 | count, no interval reported | 2026-08-24 |
Source files: hard-llm-ortier-hermes4-405b.json, hard-llm-level-hermes4-405b-statute_rag.json. Raw model name in source: nousresearch/hermes-4-405b.