Hermes 3 70 B vs Llama 3.1 70 B Instruct Fp 8: cost comparison
Prices updated 2026-10-27 · computed from live provider data
Short answer: Llama 3.1 70 B Instruct Fp 8 is cheaper on a
3:1 input/output blend — $0.165 per million tokens versus
$0.165, a 0% difference.
Both offer the same 131,072-token context window, so cost is the deciding factor.
Head to head
| Hermes 3 70 B | Llama 3.1 70 B Instruct Fp 8 | |
|---|---|---|
| Input, per 1M tokens | $0.120 | $0.120 |
| Output, per 1M tokens | $0.300 | $0.300 |
| Blended 3:1, per 1M | $0.165 | $0.165 |
| Context window | 131,072 | 131,072 |
| Vendor | lambda_ai | lambda_ai |
Monthly cost at real volumes
| Workload (input / output) | Hermes 3 70 B | Llama 3.1 70 B Instruct Fp 8 |
|---|---|---|
| 1M in / 250K out | $0.195 | $0.195 |
| 10M in / 2.5M out | $1.95 | $1.95 |
| 100M in / 25M out | $19.50 | $19.50 |
Which should you pick?
Because the context windows are identical, the cheaper option is the better default. Choose Llama 3.1 70 B Instruct Fp 8 unless you are already standardised on lambda_ai tooling and want to avoid a second integration.