DeepSeek Coder vs Gemma 4 31 B It Nvfp 4: cost comparison
Prices updated 2026-10-28 · computed from live provider data
Short answer: DeepSeek Coder is cheaper on a
3:1 input/output blend — $0.175 per million tokens versus
$0.205, a 15% difference.
Gemma 4 31 B It Nvfp 4 offers the larger context window at 262,144 tokens, against 128,000 for DeepSeek Coder.
Head to head
| DeepSeek Coder | Gemma 4 31 B It Nvfp 4 | |
|---|---|---|
| Input, per 1M tokens | $0.140 | $0.140 |
| Output, per 1M tokens | $0.280 | $0.400 |
| Blended 3:1, per 1M | $0.175 | $0.205 |
| Context window | 128,000 | 262,144 |
| Vendor | deepseek | sail |
Monthly cost at real volumes
| Workload (input / output) | DeepSeek Coder | Gemma 4 31 B It Nvfp 4 |
|---|---|---|
| 1M in / 250K out | $0.210 | $0.240 |
| 10M in / 2.5M out | $2.10 | $2.40 |
| 100M in / 25M out | $21.00 | $24.00 |
Which should you pick?
Pick Gemma 4 31 B It Nvfp 4 when you need the longer context window (262,144 vs 128,000 tokens). Pick DeepSeek Coder when price per token dominates your workload — high-volume classification, extraction, or batch summarisation.