Ad

Privum

We build, secure and run your digital infrastructure.

Software, AI systems, penetration testing and secure cloud for companies.

Talk to Privum

deepseek-ai

DeepSeek cuts output prices for V4 Flash models

412 words

DeepSeek changed OpenRouter prices for three models on 2026-11-17: DeepSeek V4 Flash 0731, DeepSeek V4.1 Flash 1017, and DeepSeek V4.1 Flash 1017 (batch).

OpenRouter price, per 1M tokens

Measured on November 17, 2026

DeepSeek: DeepSeek V4 Flash 0731
Input /M tokens
$0.06
Output /M tokens
$0.18

Price changes in the last 90 days

  1. input $0.043 → $0.06, output $0.30 → $0.18
  2. input $0.059 → $0.043, output $0.30 → $0.30
  3. input $0.119 → $0.059, output $0.238 → $0.30
  4. input $0.44 → $0.119, output $1.32 → $0.238
  5. input $0.044 → $0.44, output $0.133 → $1.32
  6. input $0.44 → $0.044, output $1.32 → $0.133

Full price history

DeepSeek: DeepSeek V4.1 Flash 1017
Input /M tokens
$0.036
Output /M tokens
$0.50

Price changes in the last 90 days

  1. input $0.02 → $0.036, output $0.108 → $0.50
  2. input $0.038 → $0.02, output $0.50 → $0.108
  3. input $0.15 → $0.038, output $0.60 → $0.50
  4. input $0.006 → $0.15, output $0.50 → $0.60
  5. input $0.15 → $0.006, output $0.60 → $0.50

Full price history

DeepSeek: DeepSeek V4.1 Flash 1017 (batch)
Input /M tokens
$0.036
Output /M tokens
$0.317

Price changes in the last 90 days

  1. input $0.029 → $0.036, output $0.253 → $0.317
  2. input $0.05 → $0.029, output $0.15 → $0.253
  3. input $0.011 → $0.05, output $0.20 → $0.15

Full price history

DeepSeek V4 Flash 0731

DeepSeek V4 Flash 0731 now costs $0.06 per million input tokens and $0.18 per million output tokens, up from $0.0434 input and $0.3 output on 2026-11-16, a 38.2 percent input increase and a 40.0 percent output cut. The model has 304.2 billion parameters, a 1048576 token context window, and open weights. Its price record runs from 2026-08-22 across 88 days with 41 price changes; the last change before today was on 2026-11-07. Today's input price of $0.06 is not a new low or high; the record shows a low of $0.0051 on 2026-10-03 and a high of $0.44 on 2026-10-20.

DeepSeek V4.1 Flash 1017 and batch variant

DeepSeek V4.1 Flash 1017 now costs $0.036 per million input tokens and $0.5 per million output tokens, up from $0.02 input and $0.108 output on 2026-11-16, an 80.0 percent input increase and a 363.0 percent output increase. The batch variant now costs $0.036 input and $0.317 output, up from $0.029 input and $0.253 output, a 24.1 percent input increase and a 25.3 percent output increase. Both have 763.2 billion parameters and a 1048576 token context window. The standard model's record spans 31 days with 5 changes, last moving on 2026-11-16; today's input price is not a new low or high, with a record low of $0.006 on 2026-11-09 and a high of $0.15 on 2026-10-18. The batch record spans 9 days with 3 changes, last moving on 2026-11-16; today's input price is not a new low or high, with a low of $0.011 on 2026-11-09 and a high of $0.05 on 2026-11-13.

DeepSeek announcement

deepseek-ai announced deepseek-ai/DeepSeek-V4-Flash-0731. According to deepseek-ai, DeepSeek-V4-Flash-0731 outperforms DeepSeek-V4-Pro (Preview) on the listed benchmarks. According to deepseek-ai, DeepSeek-V4-Flash-0731 is broadly competitive with the strongest proprietary models available. The release supersedes the preview version, shares the model structure with DeepSeek-V4-Flash-DSpark, and includes a speculative decoding module. The reasoning_effort parameter supports low, high, and max levels, with a recommended maximum output length of 384K tokens for high and max. The repository and weights are under the MIT License.

  • DeepSeek V4 Flash 0731 input price rose 38.2 percent while output price fell 40.0 percent.
  • DeepSeek V4.1 Flash 1017 input price rose 80.0 percent and output price rose 363.0 percent.
  • deepseek-ai claims DeepSeek-V4-Flash-0731 outperforms DeepSeek-V4-Pro (Preview) on the listed benchmarks.
DeepSeek cuts output prices for V4 Flash models — overfeed.news