DeepSeek V4 Flash 0731 vs Ministral 3 3B
Price and specs side by side · USD per 1M tokens · verified 2026-10-11
// side by side
| metric | DeepSeek V4 Flash 0731 | Ministral 3 3B |
|---|
| input / 1M | $0.06 | $0.10 |
| output / 1M | $0.18 | $0.10 |
| blended (3:1) | $0.09 | $0.10 |
| cache read / 1M | $0.015 | $0.01 |
| context | 1.05M | 131.1K |
| max output | 384K | 262.1K |
| input types | text | text · image |
| reasoning | yes | no |
| tool calling | yes | yes |
| structured output | yes | yes |
| open weights | yes | yes |
| released | 2026-07-31 | 2025-12-02 |
| knowledge cutoff | 2025-05 | n/a |
| lab | DeepSeek | Mistral |
// monthly cost scenarios
| workload | tokens in / out | DeepSeek V4 Flash 0731 | Ministral 3 3B |
|---|
| Chatbot | 10M / 2M | $0.96 | $1.20 |
| RAG app | 50M / 5M | $3.90 | $5.50 |
| Coding agent | 20M / 10M | $3.00 | $3.00 |
// verdict
- DeepSeek V4 Flash 0731 is 10% cheaper than Ministral 3 3B on a blended 3:1 input/output basis.
- DeepSeek V4 Flash 0731 has the larger context window (1.05M vs 131.1K tokens).
- Only DeepSeek V4 Flash 0731 supports built-in reasoning.
- Only Ministral 3 3B accepts image input.