nousresearch
Nous: Hermes 4 405B
nousresearch/hermes-4-405b
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
Hermes 4 405B is a text-only model from Nous Research, fine-tuned from Meta's Llama-3.1-405B and released under the Llama 3 license, according to its Hugging Face card. The card describes a hybrid reasoning mode with optional internal deliberation, tool calling, and schema-faithful JSON generation. OpenRouter's catalog lists a 131,072-token context window and an August 2024 knowledge cutoff. Vendor-reported gains in math, code, and steerability remain independently unverified in this source pack.[1][2]
Strengths
- Vendor-reported hybrid reasoning: the card says it can answer directly or deliberate inside think tags, toggled via a chat-template flag or system prompt, with reasoning also an optional OpenRouter parameter.[1][2]
- The card reports training for schema-faithful structured outputs, malformed-JSON repair, and function calls interleaved with reasoning; OpenRouter lists structured_outputs among supported parameters.[1][2]
- OpenRouter's catalog lists a 131,072-token context window, positioning it for long inputs.[1]
Limitations
- Performance evidence is thin: the card's benchmark tables are images without verifiable figures, its metadata results list is empty, and vendor claims such as RefusalBench leadership lack independent confirmation.[2]
- Text-only input and output according to the OpenRouter catalog; no multimodal capability is documented.[1]
- Practical gaps remain: hardware requirements, deployment cost, and real-workload behavior are not documented in this pack, so validation on your own tasks is advised before adoption.[1][2]
Where to use it
- May suit chat assistants where operators want adjustable reasoning effort, since deliberation is optional per the model card.[2]
- Potentially useful for agent-style pipelines needing tool calls or schema-conformant JSON, a vendor-reported strength worth testing against your own schemas.[1][2]
- Could be evaluated for long-document summarization or analysis given the catalog's 131k-token context, though real performance at that length is unverified.[1]
Sources
- [1] OpenRouter catalog · Nous: Hermes 4 405BSources checked: Nov 13, 2026
- [2] Model documentation · Nous: Hermes 4 405BSources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
Hermes 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...
Specifications
- Model ID
- nousresearch/hermes-4-405b
- Creator
- nousresearch
- Added to OpenRouter
- 26.08.2025
- Context length
- 131,072
- Modality
- text->text
- Input modalities
- text
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $1 / out $3
- Knowledge cutoff
- 31.08.2024
- Expiration date
- —
Tasks and capabilities
A grade appears only where the source supplies a matching measurement for that task, the model is active and the cohort holds at least 10 comparable models. The capability column lists facts the source declares — supported formats, reasoning and tool parameters — which are not measured quality.
| Task | Grade | Source-declared capability |
|---|---|---|
| General quality | No measurement | reasoning: optional · parameters: include_reasoning, reasoning |
| Coding | No measurement | — |
| Agents | No measurement | parameters: structured_outputs |
| Writing | No measurement | output: text |
| Reasoning | No measurement | reasoning: optional · parameters: include_reasoning, reasoning |
| Document analysis | No measurement | — |
| Vision | No measurement | — |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | — |
| Embeddings | No measurement | — |
Evaluations
No evaluations are supplied for this model.
How grades work
Each letter grade compares a model only against other models scored by the same source and metric in the current snapshot. S means the top 10%, A the next 20%, B the next 30%, C the rest. Models without a published measurement are shown as not rated — a supported format alone never implies measured quality.
- S — S — top 10% of the cohort
- A — A — next 20% of the cohort
- B — B — next 30% of the cohort
- C — C — below the B band
Providers
Optional live provider statistics from OpenRouter. Missing values are unknown, not zero.
| Provider | Context length | Price in /1M | Price out /1M | Latency p50 (30m, s) | Throughput p50 (30m, tok/s) | Uptime (30m) |
|---|---|---|---|---|---|---|
| Nebiusfp8 | 131,072 | $1 | $3 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.