Hermes 4 405B from Nous Research is a hybrid reasoning model with 131K context, think-tag deliberation, schema adherence, and tool use.
Capabilities, design details, and architectural traits
Hermes 4 405B is Nous Research’s frontier hybrid-mode reasoning model based on Llama-3.1-405B. Official documentation positions it around explicit deliberation, structured outputs, and controllable reasoning behavior.
| Aspect | Official Characteristic |
|---|---|
| Reasoning mode | Uses a hybrid approach that can either deliberate with <think>...</think> traces or answer directly. |
| Reasoning control | Reasoning can be toggled with thinking=True or a reasoning.enabled setting. |
| Context strategy | Supports a 131,072-token context window. |
| Structured output | Documented for JSON mode, schema adherence, and tool/function calling. |
| Training focus | Trained on an expanded post-training corpus emphasizing reasoning traces. |
| Alignment profile | Described as steerable, with lower refusal rates and neutral, user-directed behavior. |
Independent evaluations · Artificial Analysis
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Nous Research | Anthropic | Anthropic |
| Release Date | August 27, 2025 | July 24, 2026 | June 9, 2026 |
| Knowledge Cutoff | Aug 2024 | May 2026 | - |
| Context & Limits | |||
| Context Window | 131K | 1M Best Context Window | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | $1 Best Input Pricing | $5 | $10 |
| Output Pricing | $3 Best Output Pricing | $25 | $50 |
| Modalities | |||
| Inputs | text | textimage | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 8.6 | 63.1 Best Intelligence Index | 62.1 |
| Coding Index | - | 78.0 Best Coding Index | 76.5 |
| Agentic Index | - | 59.2 Best Agentic Index | 56.6 |
Graduate-level reasoning and expert Q&A evaluation.
Extremely difficult logical reasoning and knowledge.
Logical reasoning over long context windows.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
