Hermes 4 405B
Summary
Hermes 4 405B is an open-weight language model from Nousresearch, released on 26 Aug 2025. It accepts text and generates text, with a 131K-token context window and up to 118K output tokens per response. As of 19 Nov 2026, the lowest listed API price is $1 per million input tokens and $3 per million output tokens (Nebius AI Studio), across 2 providers we track.
Reasoning model for deliberate analysis, multi-step problem solving, and tool use
Last verified 19 Nov 2026. Every figure on this page links to its source.
Specifications
| Developer | Nousresearch |
|---|---|
| Release date | 26 Aug 2025 [models.dev] |
| Status | Generally available [models.dev] |
| Weights | Open weights [Hugging Face] |
| License | llama3 [Hugging Face] |
| Parameters | 406B [Hugging Face] |
| Context window | 131K (131,072 tokens) [models.dev] |
| Max output | 118K tokens [models.dev] |
| Knowledge cutoff | 31 Aug 2024 [models.dev] |
| Input | Text [models.dev] |
| Output | Text [models.dev] |
| Reasoning mode | Yes [models.dev] |
| Tool calling | No [models.dev] |
| Structured output | Yes [models.dev] |
| Hugging Face | NousResearch/Hermes-4-405B [Hugging Face] |
Hermes 4 405B API pricing by provider
USD per million tokens. Sorted by input price.
| Provider | Input | Output | Cached input | Context | Source |
|---|---|---|---|---|---|
| Nebius AI Studio NousResearch/Hermes-4-405B | $1 | $3 | $1 | 131K | models.dev, 19 Nov 2026 |
| OpenRouter nousresearch/hermes-4-405b | $1 | $3 | $0.20 | 131K | OpenRouter, 19 Nov 2026 |
Running Hermes 4 405B locally
Estimated memory, from the parameter count and the architecture in config.json on Hugging Face. Includes about 0.8 GB of runtime overhead.
| Quantization | Weights | Total, 8K context | Total, 32K context |
|---|---|---|---|
| Q4_K_M (4-bit) | 246 GB | 251 GB | 264 GB |
| Q8_0 (8-bit) | 431 GB | 436 GB | 449 GB |
| FP16 / BF16 (16-bit) | 812 GB | 817 GB | 829 GB |
At 4-bit with an 8K conversation, Hermes 4 405B needs about 251 GB, more than any single consumer graphics card; it needs a unified-memory machine, a multi-GPU setup or a split between graphics card and system RAM. Check your own hardware, or see Best computer for local AI on our sister site AIPCs.
Frequently asked questions
What is the context window of Hermes 4 405B?
Hermes 4 405B has a context window of 131,072 tokens (131K) and can generate up to 117,964 tokens in a single response, according to models.dev.
How much does the Hermes 4 405B API cost?
Hermes 4 405B is listed from $1 per million input tokens and $3 per million output tokens (Nebius AI Studio). Prices last verified 19 Nov 2026.
Is Hermes 4 405B open source?
Hermes 4 405B is an open-weight model: its weights are published on Hugging Face as NousResearch/Hermes-4-405B under the llama3 license. Check the license terms before commercial use.
When was Hermes 4 405B released?
Nousresearch released Hermes 4 405B on 26 Aug 2025, according to models.dev.
Which providers offer Hermes 4 405B?
We track Hermes 4 405B on 2 providers: Nebius AI Studio, OpenRouter.
How much VRAM does Hermes 4 405B need?
About 251 GB at 4-bit (Q4_K_M) with an 8K-token conversation, and about 436 GB at 8-bit. That is more than any single consumer graphics card holds. These are estimates from the model's published parameter count and architecture.
Similar models from other labs
Comparable on output type, weights, release date, price and context window. Prices are the lowest listed, USD per million tokens.
| Model | Lab | Released | Context | Input $/1M | Output $/1M | Weights |
|---|---|---|---|---|---|---|
| Qwen3 VL 235B A22B Thinking | Alibaba (Qwen) | 23 Sept 2025 | 131K | $0.98 | $3.95 | Open |
| GLM-4.5 | Zhipu AI (Z.ai) | 28 Jul 2025 | 131K | $0.60 | $2.20 | Open |
| Kimi K2 0711 | Moonshot AI | 11 Jul 2025 | 131K | $0.57 | $2.30 | Open |
| Kimi-K2-Instruct | Moonshot AI | 11 Jul 2025 | 131K | $0.57 | $2.30 | Open |
| Magistral Small 1.2 | Mistral AI | 18 Sept 2025 | 128K | $0.50 | $1.50 | Open |
| Qwen3-235B-A22B-Thinking-2507 | Alibaba (Qwen) | 25 Jul 2025 | 262K | $0.30 | $3 | Open |
Change history
- Max output changed from 8K to 118K
- Name changed from Hermes-4-405B to Hermes 4 405B
- Tool calling changed from yes to no
- Tool calling changed from no to yes
- Max output changed from 118K to 8K
- Name changed from Hermes 4 405B to Hermes-4-405B
- Name changed from Hermes-4-405B to Hermes 4 405B
- Tool calling changed from yes to no
- Max output changed from 8K to 118K
- Max output changed from 118K to 8K
- Tool calling changed from no to yes
- Name changed from Hermes 4 405B to Hermes-4-405B
- Max output changed from 8K to 118K
- Tool calling changed from yes to no
- Name changed from Hermes-4-405B to Hermes 4 405B
- Max output changed from 118K to 8K
- Tool calling changed from no to yes
- Name changed from Hermes 4 405B to Hermes-4-405B
- Max output changed from 8K to 118K
- Tool calling changed from yes to no
- Name changed from Hermes-4-405B to Hermes 4 405B
- Name changed from Hermes 4 405B to Hermes-4-405B
- Max output changed from 118K to 8K
- Tool calling changed from no to yes
- Tool calling changed from yes to no