Meta Llama 3.1.8 B Instruct Turbo vs Kimi Latest: cost comparison
Prices updated 2026-11-19 · computed from live provider data
Short answer: Meta Llama 3.1.8 B Instruct Turbo is cheaper on a
3:1 input/output blend — $0.180 per million tokens versus
$2.89, a 94% difference.
Kimi Latest offers the larger context window at 1,048,576 tokens, against 131,072 for Meta Llama 3.1.8 B Instruct Turbo.
Head to head
| Meta Llama 3.1.8 B Instruct Turbo | Kimi Latest | |
|---|---|---|
| Input, per 1M tokens | $0.180 | $0.182 |
| Output, per 1M tokens | $0.180 | $11.01 |
| Blended 3:1, per 1M | $0.180 | $2.89 |
| Context window | 131,072 | 1,048,576 |
| Vendor | together_ai | moonshotai |
Monthly cost at real volumes
| Workload (input / output) | Meta Llama 3.1.8 B Instruct Turbo | Kimi Latest |
|---|---|---|
| 1M in / 250K out | $0.225 | $2.93 |
| 10M in / 2.5M out | $2.25 | $29.34 |
| 100M in / 25M out | $22.50 | $293 |
Which should you pick?
Pick Kimi Latest when you need the longer context window (1,048,576 vs 131,072 tokens). Pick Meta Llama 3.1.8 B Instruct Turbo when price per token dominates your workload — high-volume classification, extraction, or batch summarisation.