DeepSeek V4 Flash 0731

DeepSeek V4 Flash 0731 represents the official release of the Flash line, succeeding an earlier preview with substantially enhanced agentic capabilities. It is built as a 284B-parameter Mixture-of-Experts transformer that activates only 13B parameters per pass, an efficiency-oriented design aimed at keeping inference costs low while preserving the reasoning depth expected of larger DeepSeek models. The model is positioned around coding, terminal-style tool use, and multi-step automation, reflecting DeepSeek's broader emphasis on practical, agent-driven tasks rather than general chat alone.

The model's open-weight release under an MIT-style license makes it attractive for teams that want to self-host or fine-tune locally, with LM Studio noting a roughly 156 GB memory footprint for the smallest deployment. A 1,048,576-token context window supports long-horizon coding sessions and extended agent traces, while confirmed support for reasoning and function calling lets developers build structured tool pipelines. DeepSeek reports that this official Flash release outperforms both the Flash preview and the V4 Pro preview on its coding and agentic benchmark suite, signaling that the smaller active parameter count did not come at the cost of capability in its target domains.

Volcengine Arkdeepseek-v4-flash-ga-260731deepseek-flash

Open provider page Provider docs

Quick Info

Provider: Volcengine Ark
Model key: deepseek-v4-flash-ga-260731
Release date: Jul 31, 2026
Last updated: Jul 31, 2026
Knowledge cutoff: 2025-05
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $0.4453
Output token cost: $1.3359

Limits

Output tokens: 384,000 tokens
Context window: 1,000,000 tokens

Transparent token rates

Compare DeepSeek V4 Flash 0731 pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about DeepSeek V4 Flash 0731

Volcengine Ark

CoverageBenchmark

Morph's third-party reference guide confirms DeepSeek V4 Flash 0731 reached general availability on July 31, 2026, with 284B total / 13B active parameters under an MIT license and a 1M-token context window (384K max output). The guide lists the first-party API model name as `deepseek-v4-flash` with image input availabl The guide situates V4 Flash 0731 alongside the larger V4-Pro-0813 sibling (1.6T/49B, MIT, same 1M/384K context) within the open-weight V4 MoE family, both verified as of September 7, 2026. It enumerates deployment surfaces including the first-party API (api.deepseek.com with OpenAI ChatCompletions, OpenAI Responses, an

Volcengine Ark

Coverage

This Medium article directly covers the July 31, 2026 GA release of DeepSeek-V4-Flash-0731, describing it as the production release of DeepSeek's fast flagship language model. The article details the architecture: 284 billion total parameters with only 13 billion activated per token via Mixture-of-Experts (MoE), suppor The article notes that DeepSeek-V4-Flash-0731 maintains the same architecture, API endpoint, latency profile, and pricing as its predecessor while delivering substantially better performance. This positions the 0731 release as a meaningful GA milestone in the V4 Flash family, emphasizing that improving model intelligen

Volcengine Ark

CoverageAnalysis

Artificial Analysis reports that DeepSeek V4 Flash 0731, released on July 31, 2026, scores 50 on its Intelligence Index — a 10-point jump over the prior DeepSeek V4 Flash (April 2026) and placing it 6 points above DeepSeek V4 Pro and within 1 point of GLM-5.2 (max, 51). The model retains the 1M-token context window and The model shows notable agentic and hallucination improvements: GDPval-AA v2 Elo rises to 1559 (from 1189), Terminal-Bench 2.1 climbs 17 points to 79%, and τ³-Bench Banking rises 8 points to 31%. AA-Omniscience Index improves by +7 to -16, driven entirely by a reduced hallucination rate (84%, down 12 points) rather tha

Volcengine Ark

Coverage

DeepSeek's official API Change Log, dated 2026-09-10, announces the release of DeepSeek-V4.1-Flash and states that the previous-generation models V4 Flash and V4 Flash Vision Exp have been retired. For compatibility, the model names deepseek-v4-flash and deepseek-v4-flash-vision-exp are temporarily routed to V4.1 Flash Because the Volcengine Ark subject 'DeepSeek V4 Flash 0731' sits in the deepseek-v4-flash lineage, this upstream retirement/routing notice is the most directly relevant first-party signal: the model name the Ark SKU maps to is now aliased to V4.1 Flash, accompanied by an API price reduction described in the same change