LONG MEMORY · 雪国
DeepSeek
5 models from DeepSeek, living in Snow Country.
Visit the DeepSeek village in 3D
| Model | Context | Input | Output | Accepts |
|---|---|---|---|---|
| DeepSeek V4.1 Flash 1017deepseek/deepseek-v4.1-flash-1017 | 1.05M | $0.150 | $0.600 | text, images |
| DeepSeek V4.1 Flashdeepseek/deepseek-v4.1-flash | 1.05M | $0.020 | $1.00 | text, images |
| DeepSeek V4 Flash Vision Expdeepseek/deepseek-v4-flash-vision-exp | 1.05M | $0.220 | $0.650 | text, images |
| DeepSeek V4 Flash 0731deepseek/deepseek-v4-flash-0731 | 1.05M | $0.040 | $0.130 | text |
| DeepSeek V4 Flash 0423deepseek/deepseek-v4-flash | 1.05M | $0.090 | $0.180 | text |
Questions and answers
Which DeepSeek model has the largest context window?
DeepSeek V4 Flash 0423 (DeepSeek), with 1.05M tokens.
Which DeepSeek model is cheapest per input token?
DeepSeek V4.1 Flash, at $0.020 per million input tokens ($1.00 per million output tokens).
Which DeepSeek models accept images?
DeepSeek V4.1 Flash 1017, DeepSeek V4.1 Flash, DeepSeek V4 Flash Vision Exp (3 of 5).
Which DeepSeek models support reasoning?
DeepSeek V4.1 Flash 1017, DeepSeek V4.1 Flash, DeepSeek V4 Flash Vision Exp, DeepSeek V4 Flash 0731, DeepSeek V4 Flash 0423 (5 of 5).
Answers are computed from the 2026-10-24 catalog snapshot.