Hermes 4 - Llama-3.1 70B (Reasoning) vs Hermes 4 - Llama-3.1 405B (Reasoning)
Comparing 2 AI models · 5 benchmarks · Nous Research
Composite Indices
Intelligence, Coding, Math
Standard Benchmarks
Academic and industry benchmarks
Benchmark Winners
Hermes 4 - Llama-3.1 70B (Reasoning)
No clear wins
Hermes 4 - Llama-3.1 405B (Reasoning)
- GPQA
- MMLU Pro
- HLE
- LiveCodeBench
- AIME 2025
| Metric | No Hermes 4 - Llama-3.1 70B (Reasoning) | No Hermes 4 - Llama-3.1 405B (Reasoning) |
|---|---|---|
| Pricing Per 1M tokens | ||
| Input Cost | $0.13/1M | $1.00/1M |
| Output Cost | $0.40/1M | $3.00/1M |
| Blended Cost 3:1 input/output ratio | $0.20/1M | $1.50/1M |
| Specifications | ||
| Organization Model creator | Nous Research | Nous Research |
| Release Date Launch date | Aug 27, 2025 | Aug 27, 2025 |
| Performance & Speed | ||
| Throughput Output speed | 76.4 tok/s | 36.0 tok/s |
| Time to First Token (TTFT) Initial response delay | 641ms | 756ms |
| Latency Time to first answer token | 26810ms | 56268ms |
| Composite Indices | ||
| Intelligence Index Overall reasoning capability | 39.2 | 41.6 |
| Coding Index Programming ability | 34.6 | 34.8 |
| Math Index Mathematical reasoning | 68.7 | 69.7 |
| Standard Benchmarks | ||
| GPQA Graduate-level reasoning | 69.9% | 72.7% |
| MMLU Pro Advanced knowledge | 81.1% | 82.9% |
| HLE Hard language evaluation | 7.9% | 10.3% |
| LiveCodeBench Real-world coding tasks | 65.3% | 68.6% |
| MATH 500 Mathematical problems | — | — |
| AIME 2025 Advanced math competition | 68.7% | 69.7% |
| AIME (Original) Math olympiad problems | — | — |
| SciCode Scientific code generation | 34.1% | 25.2% |
| LCR Code review capability | 6.7% | 20.7% |
| IFBench Instruction-following | 31.3% | 32.7% |
| TAU-bench v2 Tool use & agentic tasks | 22.5% | 22.2% |
| TerminalBench Hard CLI command generation | 4.3% | 10.6% |
Key Takeaways
Hermes 4 - Llama-3.1 70B (Reasoning) offers the best value at $0.13/1M, making it ideal for high-volume applications and cost-conscious projects.
Hermes 4 - Llama-3.1 405B (Reasoning) leads in reasoning capabilities with a 72.7% GPQA score, excelling at complex analytical tasks and problem-solving.
Hermes 4 - Llama-3.1 405B (Reasoning) achieves a 34.8 coding index, making it the top choice for software development and code generation tasks.
All models support context windows of ∞+ tokens, suitable for processing lengthy documents and maintaining extended conversations.
When to Choose Each Model
Hermes 4 - Llama-3.1 70B (Reasoning)
- Cost-sensitive applications
- High-volume processing
Hermes 4 - Llama-3.1 405B (Reasoning)
- Complex reasoning tasks
- Research & analysis
- Code generation
- Software development
Try Models for Free
Try Hermes 4 - Llama-3.1 70B (Reasoning) for FREE
No credit card or account required.
Try Hermes 4 - Llama-3.1 405B (Reasoning) for FREE
No credit card or account required.
Cost Calculator
Hermes 4 - Llama-3.1 70B (Reasoning) Hermes 4 - Llama-3.1 405B (Reasoning) Costs are estimates based on API pricing. Actual costs may vary based on caching, batch processing, and volume discounts.
Recommended Comparisons
