DeepSeek-V4-Flash API enters public beta with agentic gains
DeepSeek released the official DeepSeek-V4-Flash API in public beta, with the model name set to deepseek-v4-flash. The update significantly enhances agent capabilities, with benchmark scores far exceeding V4-Pro-Preview, and natively supports the Responses API format with Codex adaptation. The model structure and size remain consistent with the preview version, with only post-training changes.
Coverage timeline
DeepSeek API Updates
DeepSeek-V4-Flash 更新 DeepSeek-V4-Flash 正式版 API 上线公测，API 调用方式不变，模型名设置为 deepseek-v4-flash 即可使用最新版本。 Agent 能力大幅增强，基准测试远超 V4-Pro-Preview： Terminal Bench 2.1: 82.7 NL2Repo: 54.2 Cybergym: 76.7 DeepSWE: 54.4 Toolathlon verified: 70.3 Agent Last Exam: 25.2 Automation Bench (Public): 25.1 DSBench-FullStack: 68.7 DSBench-Hard: 59.6 注1：对于公开基准测试集中的 Code Agent 任务，正式版 DeepSeek-V4-Flash 使用 DeepSeek Harness 极简模式（即将发布）作为框架进行测试，并使用 max 档位，topp=0.95，temperature=1.0 注2：DSBench-FullStack 是内部使用的全栈开发测试集，DSBench-Hard 是内部使用的 Coding Agent 难题测试集 正式版 V4-Flash 原生支持 Responses API 格式并针对性适配 Codex，具体配置方法请参考 文档 DeepSeek-V4-Flash-0731 的模型结构、尺寸和 DeepSeek-V4-Flash-Preview 保持一致，仅重新进行了后训练。 注意：本次仅升级了 DeepSeek-V4-Flash 的 API 接口，DeepSeek-V4-Pro API 及 APP/WEB 端模型未做更改。 DeepSeek-V4-Pro 正式版将会尽快发布。
* * * ## Date: 2026-07-31 ### DeepSeek-V4-Flash Update The official release of the DeepSeek-V4-Flash API is now in public beta. The API calling method remains unchanged — simply set the model name to `deepseek-v4-flash` to use the latest version. **Significantly enhanced agent capabilities, with benchmark results far exceeding V4-Pro-Preview:** * Terminal Bench 2.1: 82.7 * NL2Repo: 54.2 * Cybergym: 76.7 * DeepSWE: 54.4 * Toolathlon verified: 70.3 * Agent Last Exam: 25.2 * Automation Bench (Public): 25.1 * DSBench-FullStack: 68.7 * DSBench-Hard: 59.6 Note 1: For the Code Agent tasks in the public benchmark sets, the official DeepSeek-V4-Flash was tested using the DeepSeek Harness minimal mode (to be released soon) as the framework, with the max effort level, topp=0.95, and temperature=1.0 Note 2: DSBench-FullStack is an internal full-stack development test set, and DSBench-Hard is an internal Coding Agent hard-problem test set **The official V4-Flash natively supports the Responses AP
Hacker Newstheanonymousone
## Intelligence ### Artificial Analysis Intelligence Index Artificial Analysis Intelligence Index v4.1.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR Reasoning models are indicated by a lightbulb icon Artificial Analysis Intelligence Index v4.1.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. SeeIntelligence Index methodologyfor further details, including a breakdown of each evaluation and how we run them. ### Artificial Analysis Intelligence Index by Open Weights / Proprietary Artificial Analysis Intelligence Index v4.1.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR Reasoning models are indicated by a lightbulb icon Artificial Analysis Intelligence Index v4.1.1 includes: GDPval-AA v2, 𝜏³
Simon Willison
deepseek-ai/DeepSeek-V4-Flash-0731 The latest release in DeepSeek's V4 family, "with substantially enhanced agentic capabilities". It's 304 billion parameters - 167GB on Hugging Face - but it appears to punch well above its weight. Artificial Analysis rank it ahead of MiniMax M3 - a 428B model. It's $0.14/million input and $0.27/million output pricing means this may currently be the best value-per-intelligence model out there. It's looking very good on the Intelligence Index vs. Cost per Intelligence Index Task chart: I got a disappointing pelican from it using the default reasoning level via OpenRouter: But when I bumped reasoning level up to high I got something much better : llm -m openrouter/deepseek/deepseek-v4-flash-0731 -t pelican -o reasoning_effort high Via Hacker News Tags: ai , generative-ai , llms , pelican-riding-a-bicycle , deepseek , llm-release , openrouter , ai-in-china , artificial-analysis
Hacker Newszhoutong
## DeepSeek V4 Flash on a single AMD MI300X This repository contains the configuration and patches I use to run `deepseek-ai/DeepSeek-V4-Flash-0731` on **one AMD MI300X** in production. It includes the Docker Compose stack, SHA-256-pinned file overlays, reference diffs against upstream, and tuning tables. The checkpoint runs as shipped, without additional weight quantization or offload. Results from the pinned stack (vLLM ROCm nightly `0.26.1rc1.dev229+g124154a88.rocm723`, AITER `0.1.19`): | Metric | Result | | --- | --- | | Single-stream decode (median per-stream, DSpark-7) | **168.6 tok/s** | | Prefill with tuned kernels | **≈ 7.9–8.5K tok/s** (6,988–7,019 tok/s on fresh prompts in the shipping profile) | | 8 concurrent streams | 542 tok/s aggregate, 90.3 tok/s median per stream | | 64-stream burst | 830 tok/s aggregate, no OOM, no engine errors | | Context | 256K validated (the architecture supports 1M) | | Weights in HBM | 156.67 GiB — **no additional quantization or weight offload