DeepSeek V4 Flash 0731 (batch)
DeepSeek · Released Jul 2026
A sparse mixture-of-experts model from DeepSeek with 13 billion active parameters out of 284 billion total, optimized for batch processing with a 1-million-token context window.
- Strengths
- Delivers frontier-class performance on coding, reasoning, and agent workflows while maintaining the efficiency of a smaller active parameter count.
- Best for
- Batch workloads involving code generation, multi-step reasoning, or autonomous task execution where latency is not a constraint.
- Limitations
- Batch-only endpoint means it is not suitable for real-time or streaming applications; for general-purpose use cases, the full-size DeepSeek V4 Pro variants may be more appropriate.
Input / 1M
$0.11
Output / 1M
$0.33
Cached input / 1M
$0.0035
Context window
1.05M
Price history
Input (solid)Output (dashed)
Price change
- 30d
- in decreased 21.4% out increased 17.9%
- 90d
- in decreased 21.4% out increased 17.9%
- 1y
- in decreased 21.4% out increased 17.9%
- Since launch
- in decreased 21.4% out increased 17.9%
Snapshots
| Effective | Input | Output | Cached in | Note | Source |
|---|---|---|---|---|---|
| 8 Sep 2026 | $0.11 | $0.33 | $0.0035 | Imported from OpenRouter | openrouter.ai |
| 29 Aug 2026 | $0.14 | $0.28 | $0.03 | Imported from OpenRouter | openrouter.ai |