Side by side pricing and context for DeepSeek V3 (Together) and DeepSeek V3 (Fireworks). Rates are curated catalog values with last verified dates on each model page.
DeepSeek V3 (Together)
DeepSeek V3 served on Together when you want DeepSeek-class quality through Together’s OpenAI-compatible API.
Input $1.25 / Output $1.25 per 1M
Context 128K · Approx
DeepSeek V3 (Fireworks)
DeepSeek V3 on Fireworks for low-latency DeepSeek-class chat. Compare with native DeepSeek API rates on the same prompt.
Input $0.900 / Output $0.900 per 1M
Context 128K · Approx
Quick take
- DeepSeek V3 (Fireworks) is cheaper on input in this catalog snapshot.
- DeepSeek V3 (Fireworks) is cheaper on output in this catalog snapshot.
- Both last verified on their model pages (2026-08-02 / 2026-08-02).
Rate table
| Metric | DeepSeek V3 (Together) | DeepSeek V3 (Fireworks) |
|---|---|---|
| Input / 1M | $1.25 | $0.900 |
| Output / 1M | $1.25 | $0.900 |
| Cached input / 1M | n/a | n/a |
| Context | 128K | 128K |
| Tokenizer | Approx | Approx |