Dolphin 3.8 B vs DeepSeek R1.7 B Qwen: cost comparison
Prices updated 2026-11-17 · computed from live provider data
Short answer: DeepSeek R1.7 B Qwen is cheaper on a
3:1 input/output blend — $0.098 per million tokens versus
$0.098, a 0% difference.
DeepSeek R1.7 B Qwen offers the larger context window at 131,072 tokens, against 128,000 for Dolphin 3.8 B.
Head to head
| Dolphin 3.8 B | DeepSeek R1.7 B Qwen | |
|---|---|---|
| Input, per 1M tokens | $0.080 | $0.080 |
| Output, per 1M tokens | $0.150 | $0.150 |
| Blended 3:1, per 1M | $0.098 | $0.098 |
| Context window | 128,000 | 131,072 |
| Vendor | llamagate | llamagate |
Monthly cost at real volumes
| Workload (input / output) | Dolphin 3.8 B | DeepSeek R1.7 B Qwen |
|---|---|---|
| 1M in / 250K out | $0.117 | $0.117 |
| 10M in / 2.5M out | $1.18 | $1.18 |
| 100M in / 25M out | $11.75 | $11.75 |
Which should you pick?
Pick DeepSeek R1.7 B Qwen when you need the longer context window (131,072 vs 128,000 tokens). Pick DeepSeek R1.7 B Qwen when price per token dominates your workload — high-volume classification, extraction, or batch summarisation.