Infrastructure
Latency and price differences for the same model across providers. Sample snapshot only.
DeepSeek price comparison across aggregators
Input/output price per 1M tokens for the same model on different platforms (sample snapshot).
Same-model comparison · DeepSeek V3
| Metric | OpenRouter | Together | SiliconFlow | Fireworks | Groq |
|---|---|---|---|---|---|
| Input / 1M | $0.14 | $0.18 | $0.12 Lowest | $0.16 | $0.22 |
| Output / 1M | $0.28 | $0.32 | $0.24 Lowest | $0.30 | $0.35 |
-
OpenRouter
- Input / 1M
- $0.14
- Output / 1M
- $0.28
-
Together
- Input / 1M
- $0.18
- Output / 1M
- $0.32
-
SiliconFlow
- Input / 1M
- $0.12 (Lowest)
- Output / 1M
- $0.24 (Lowest)
-
Fireworks
- Input / 1M
- $0.16
- Output / 1M
- $0.30
-
Groq
- Input / 1M
- $0.22
- Output / 1M
- $0.35
Card view on narrow screens; full comparison table on large screens.
Same-model comparison · DeepSeek V3
Fastest · Fireworks
280ms
P50 latency
Fastest
Best price · OpenRouter
$0.14 / $0.28
Input / 1M · Output / 1M
Best price
Same-model comparison
-
OpenRouter
Best price- P50 latency
- 420ms
- Input / 1M
- $0.14
- Output / 1M
- $0.28
-
Together
- P50 latency
- 310ms
- Input / 1M
- $0.18
- Output / 1M
- $0.32
-
SiliconFlow
Domestic payment- P50 latency
- 580ms
- Input / 1M
- $0.12
- Output / 1M
- $0.24
-
Fireworks
Fastest- P50 latency
- 280ms
- Input / 1M
- $0.16
- Output / 1M
- $0.30