DeepSeek V4 Flash
Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work
- Strongest at
- Writing Better than 38% of the past year's models · 3 of 4 benchmarks
- Price per 1M tokens
- $0.14 / $0.28 input / output · OpenRouter
- Context window
- 1M tokens in · 384K out
- Since the last one
- 1 days after DeepSeek V4 Flash 0423
Capability profile
Scored against 91 models released in the past year. 50 is the average model; further out is better.
No results yet: Visual reasoning
Where these scores come from
Each capability combines the published benchmarks listed here, rescaled so 50 is the reference average and 15 points is one standard deviation. A hollow dot rests on a single benchmark.
- Coding LiveBench: Coding + Agentic Coding
- Writing EQ-Bench Creative Writing v3, EQ-Bench Longform Writing, LiveBench: Instruction Following
- Reading & documents LiveBench: Language + Data Analysis
- Reasoning LiveBench: Reasoning
- Math LiveBench: Mathematics
- Science WeirdML v2
- Computer use AutomationBench 1.0.6
Price against capability
The shaded corner holds anything better and cheaper. Prices blend input and output 3 to 1.
Nothing from the past year scores higher on Writing for less.
1 model from the past year scores higher on Coding for less: DeepSeek V4 Flash 0731.
1 model from the past year scores higher on Reading & documents for less: DeepSeek V4 Flash 0731.
2 models from the past year score higher on Reasoning for less: DeepSeek V4 Flash 0731, Qwen3.7 Flash.
2 models from the past year score higher on Math for less: DeepSeek V4 Flash 0731, Qwen3.7 Flash.
2 models from the past year score higher on Science for less: DeepSeek V4 Flash 0731, Qwen3.7 Flash.
Nothing from the past year scores higher on Computer use for less.
OpenRouter $0.14 / $0.28 · Cached input $0.028
Since DeepSeek V4 Flash 0423
5 releases in the DeepSeek line.
A newer release is out: DeepSeek V4.1 Flash 1017.
Epoch benchmark boards
Best result on each board it has run. The tick marks the leader.
DeepSeek's past year
9 release days in the past year, about one every 41 days. The next is expected around Nov 28, 2026. Forecasts ↗
Specifications
Specifications and prices from Models.dev; availability and limits may differ by provider. Benchmarks from Epoch AI (CC BY 4.0), LiveBench and the other sources named above.