Mercury 2.5 vs Llama 3.1.8 B Instant: cost comparison
Prices updated 2026-11-09 · computed from live provider data
Short answer: Llama 3.1.8 B Instant is cheaper on a
3:1 input/output blend — $0.275 per million tokens versus
$0.338, a 19% difference.
Mercury 2.5 offers the larger context window at 260,000 tokens, against 131,072 for Llama 3.1.8 B Instant.
Head to head
| Mercury 2.5 | Llama 3.1.8 B Instant | |
|---|---|---|
| Input, per 1M tokens | $0.200 | $0.200 |
| Output, per 1M tokens | $0.750 | $0.500 |
| Blended 3:1, per 1M | $0.338 | $0.275 |
| Context window | 260,000 | 131,072 |
| Vendor | inception | inferenceapis |
Monthly cost at real volumes
| Workload (input / output) | Mercury 2.5 | Llama 3.1.8 B Instant |
|---|---|---|
| 1M in / 250K out | $0.388 | $0.325 |
| 10M in / 2.5M out | $3.88 | $3.25 |
| 100M in / 25M out | $38.75 | $32.50 |
Which should you pick?
Pick Mercury 2.5 when you need the longer context window (260,000 vs 131,072 tokens). Pick Llama 3.1.8 B Instant when price per token dominates your workload — high-volume classification, extraction, or batch summarisation.