Back to Explore
DeepSeek: DeepSeek V4 Flash 0731
DeepSeek
Public weightsTrending304B1.0M ctxMIT
Comparable data available52% coverage14/14 fields directly observedUpdated Jul 31, 2026, 6:15 PM
About
DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.
Related Models
DeepSeek V4 Flash 0731 (Reasoning, Max Effort)2026-07-31DeepSeek V4 Pro (Reasoning, Max Effort)2026-04-24DeepSeek V4 Pro (Reasoning, High Effort)2026-04-24DeepSeek V4 Flash (Reasoning, Max Effort)2026-04-24DeepSeek V4 Flash (Reasoning, High Effort)2026-04-24DeepSeek V4 Pro (Non-reasoning)2026-04-24DeepSeek V4 Flash (Non-reasoning)2026-04-24DeepSeek V3.2 (Reasoning)2025-12-01
Pricing
Input
$0.14
per 1M tokens
Output
$0.28
per 1M tokens
Blended
$0.17
per 1M tokens
Cheaper than 82% of models. Median price is $0.85/1M tokens.
Cost Calculator
Tokens per day1M
100K100M
Daily
$0.17
Monthly
$5.25
vs. Similar Models
DeepSeek V4 Flash 0731 (Reasoning, Max Effort)
$0.170%
DeepSeek V4 Flash (Reasoning, Max Effort)
$0.170%
DeepSeek V4 Flash (Reasoning, High Effort)
$0.170%
MiMo-V2.5
$0.170%
Performance
Context Window
1.0M
tokens
Larger than 86% of models
Max Output
384K
tokens
37% of context
Context Window Comparison
Kimi K3 (max)
1.0MSame
GLM-5.2 (max)
1.0MSame
Muse Spark 1.1 (xhigh)
1.0MSame