Nous Research model product
Hermes 3 70B
Hermes 3 70B is Nous Research's flagship instruction-following model, fine-tuned for advanced reasoning, creative writing, and complex task completion.
Updated Aug 12, 2026. Default version: Hermes 3 70B
LLMBoard score7.2Hermes 3 70B
Coverage80%16 benchmark families
Context window131.1KTokens
Official input priceN/AOfficial price unavailable
Hermes 3 70B Specifications
Technical details for the model's default version.
- Version
- Hermes 3 70B
- Released
- Aug 15, 2024
- Knowledge cutoff
- Unknown
- Parameters
- 70B
- Context window
- 131.1K
- Max output
- 16.4K
- Inputs
- text
- Outputs
- text
- Open weights
- No
- License
- Apache 2.0
Hermes 3 70B Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Hermes 3 70B Benchmark Results
Benchmark scores for Hermes 3 70B.
| BoolQ | 88.0% | 01 | 10 | 100.0% | C | |
| MuSR | 50.7% | 02 | 2 | 0.0% | C | |
| PIQA | 84.4% | 02 | 11 | 90.0% | C | |
| ARC-E | 83.0% | 03 | 8 | 71.4% | C | |
| IFBench | 81.2% | 03 | 29 | 92.9% | C | |
| AGIEval | 56.2% | 04 | 10 | 66.7% | C | |
| MT-Bench | 89.9% | 04 | 12 | 72.7% | C | |
| OpenBookQA | 49.4% | 05 | 5 | 0.0% | C | |
| TruthfulQA | 63.3% | 06 | 18 | 70.6% | C | |
| HellaSwag | 88.2% | 07 | 27 | 76.9% | C | |
| Winogrande | 83.2% | 07 | 22 | 71.4% | C | |
| BBH | 67.8% | 11 | 12 | 9.1% | C | |
| ARC-C | 65.5% | 27 | 34 | 21.2% | C | |
| MMLU | 79.1% | 66 | 100 | 34.3% | C | |
| MATH | 20.8% | 70 | 71 | 1.4% | C | |
| MMLU-Pro | 47.2% | 118 | 129 | 8.6% | C | |
| GPQA | 66.1% | 145 | 234 | 38.2% | C |
Hermes 3 70B Arena Results
Preference and agent-evaluation results for the default version.
No Arena results
The default version does not have a matching Arena result yet.
Hermes 3 70B Runtime Performance
Provider-specific output speed and catalog latency for Hermes 3 70B. Runtime does not affect the capability score.
No runtime data
Browse runtime rankingsNo provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Hermes 3 70B Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- N/A
- Official provider
- N/A
- Lowest third-party
- From $0.408 input, $0.408 output per 1M via NanoGPT
- Tracked offerings
- 1
| NanoGPT | NousResearch/hermes-3-llama-3.1-70b | global | $0.408 | $0.408 | 65.5K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Hermes 3 70B Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
| Hermes 3 70B | 7.2 | 70B | 131.1K | 16.4K | No | Apache 2.0 |
What is Hermes 3 70B?
Key information about Hermes 3 70B and its available data.
Hermes 3 70B is Nous Research's flagship instruction-following model, fine-tuned for advanced reasoning, creative writing, and complex task completion. It features exceptional instruction adherence and strong performance across multiple domains.
Data as of 2026-08-11.
Hermes 3 70B vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to Hermes 3 70B
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
MI
Phi 4 multimodal
Microsoft
MA
Mistral Small 3 24B
Mistral AI
MA
Mistral Large 2
Mistral AI
AC
Qwen2.5 7B
Alibaba Cloud / Qwen Team
OP
GPT-4o-mini
OpenAI
AC
Qwen2.5 14B
Alibaba Cloud / Qwen Team
FAQ
Common questions about Hermes 3 70B.
When was Hermes 3 70B released?
Hermes 3 70B's default version was released on Aug 15, 2024.
How much does Hermes 3 70B cost?
No official standard PAYG price is currently available for Hermes 3 70B. The lowest tracked third-party offer starts at $0.408 input and $0.408 output via NanoGPT.
Who created Hermes 3 70B?
Hermes 3 70B was created by Nous Research.
What is the context window for Hermes 3 70B?
The default version has a 131.1K token context window.
Is Hermes 3 70B open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer Hermes 3 70B?
1 provider offerings are linked to the default version.
What models should I compare Hermes 3 70B with?
Nearby ranked alternatives include GPT-4o-mini, Mistral Small 3 24B, Phi 4 multimodal.