deepseek-v4-flash-0731 vs deepseek-v4.1-flash
| deepseek-v4-flash-0731 | deepseek-v4.1-flash | |
|---|---|---|
| Providers | DeepSeek | DeepSeek |
| Input $/M | $0.15 | $0.15 |
| Output $/M | $0.60 | $0.60 |
| Output multiplier | ×4.00 | ×4.00 |
| Context window | 1M tokens | 1M tokens |
| Intelligence index | 34.3 | 39.5 |
| Coding index | 69.1 | — |
What the numbers mean
Input pricing is close — 1.00× apart — so the decision should come down to measured quality on your own evaluation set rather than headline cost.
On measured intelligence index the gap is 5.2 points, with deepseek-v4.1-flash ahead. That is a large enough spread to be visible on multi-step tasks.
Tags: deepseek-v4-flash-0731 — Chat, Vision, Reasoning, Coding, 工具调用, Long context, 超大输出, 快速 · deepseek-v4.1-flash — Chat, Vision, Reasoning, Coding, 工具调用, Long context, 超大输出, 快速
Code
from openai import OpenAI
client = OpenAI(base_url="https://starseaapi.com/v1", api_key="sk-...")
# deepseek-v4-flash-0731
client.chat.completions.create(model="deepseek-v4-flash-0731", messages=[{"role":"user","content":"Hi"}])
# deepseek-v4.1-flash
client.chat.completions.create(model="deepseek-v4.1-flash", messages=[{"role": "user", "content": "Hi"}])