DeepSeek V4-Pro
PulseAugur coverage of DeepSeek V4-Pro — every cluster mentioning DeepSeek V4-Pro across labs, papers, and developer communities, ranked by signal.
- uses DeepClaude 90%
- instance of DeepSeek V4-Flash 90%
- used by Tencent 90%
- used by Huawei Ascend 90%
- instance of MIT License 90%
- used by Fireworks AI 90%
- affiliated with Fireworks AI 70%
- used by MIT License 70%
- competes with DeepSeek V4-Flash 60%
- affiliated with DeepClaude 60%
- competes with Opus 4.7 60%
- used by Opus 4.7 50%
- 2026-05-25 product_launch DeepSeek has reportedly launched its V4 Pro model. 来源
- 2026-05-25 product_launch DeepSeek V4-Pro has reduced its pricing by 75%. 来源
- 2026-05-25 product_launch DeepSeek V4 Pro has been released with a 75% price reduction. 来源
- 2026-05-24 product_launch DeepSeek permanently reduced the price of its V4 Pro model. 来源
- 2026-05-24 product_launch DeepSeek V4 Pro has been recognized as the global leader in cost-efficiency following a permanent price reduction.
- 2026-05-23 product_launch DeepSeek permanently reduced the price of its V4 Pro model by 75%.
- 2026-05-23 product_launch DeepSeek permanently reduced the price of its V4 Pro model by 75%.
- 2026-05-23 product_launch DeepSeek-V4-Pro API pricing will adjust after a promotional period. 来源
- 2026-05-22 product_launch DeepSeek permanently reduced the API price of its DeepSeek V4 Pro model by 75%.
- 2026-05-22 product_launch DeepSeek V4 Pro API model pricing permanently reduced by 75%.
- 2026-05-22 product_launch DeepSeek V4 Pro API pricing is permanently reduced to 25% of its original price after a promotional period.
- 2026-05-19 product_launch DeepSeek v4 Pro is being used as a competitive alternative to Claude Opus for various AI tasks. 来源
- 2026-05-12 research_milestone NIST's CAISI evaluated DeepSeek V4 Pro, finding it comparable to GPT-5 and the top-performing Chinese AI model. 来源
- 2026-04-29 product_launch DeepSeek-V4 Pro model with long-context capabilities is now available on the Together AI platform. 来源
- 2026-04-27 product_launch Fireworks AI launched the DeepSeek V4 Pro model after addressing critical bugs. 来源
16 天有情绪数据
-
小型模型在代理编码基准测试中超越前沿AI
一项最近的代理编码基准测试显示，更小、更高效的模型在性能上超越了更大、更前沿的模型。SmolLM3 3B 模型能够在笔记本电脑上运行，得分达到 93.3，显著超过了 Grok 4.20 和 DeepSeek V4 Pro 等模型。这表明模型大小可能不是代理编码能力的决定性因素，挑战了之前关于高级任务必须拥有海量参数的假设。
-
NIST: DeepSeek V4 Pro matches GPT-5 performance, leads China models
The U.S. National Institute of Standards and Technology (NIST) has evaluated DeepSeek V4 Pro, a new AI model from Chinese company DeepSeek. The evaluation found that DeepSeek V4 Pro performs comparably to OpenAI's GPT-5…
-
腾讯发布 Hy-MT2 多语言翻译模型系列
腾讯发布了其 Hy-MT2 系列多语言翻译模型，提供 1.8B、7B 和 30B-A3B 等不同规模。这些模型支持 33 种语言之间的翻译，专为复杂的真实世界场景设计，包括指令遵循。1.8B 模型采用极致量化技术，可在设备上部署，大小缩减至 440MB，同时提升推理速度。Hy-MT2 模型表现强劲，其中 7B 和 30B-A3B 版本在性能上超越了 DeepSeek-V4-Pro 和 Kimi K2.6 等开源竞争对手，而 1.8B …
-
DeepSeek V4 benchmarks show 85 tok/s at 524k context; Ollama guide for Ryzen APUs released
New benchmarks reveal DeepSeek V4 Flash achieving 85 tokens per second with a 524k context window, utilizing MTP self-speculation and FP8 quantization on dual RTX PRO 6000 Max-Q GPUs. Additionally, a guide has been publ…
-
Baidu's Wenxin 5.1 leads China in search, slashes training costs
Baidu has released its new large language model, Wenxin 5.1, which significantly enhances search, knowledge, and AI agent capabilities. The model achieves leading domestic search performance and surpasses DeepSeek-V4-Pr…
-
LLM routers struggle with rate limits and response format drift
A recent analysis highlights two critical failure modes in multi-provider LLM routing systems that can lead to unexpected costs and downtime. One issue involves how routers incorrectly handle rate limit errors, applying…
-
New research reveals universal adversarial attacks on VLMs are less effective than previously thought
Researchers have developed a new evaluation method, VisInject, to distinguish between general disruption and precise injection in adversarial attacks on vision-language models. Their findings indicate that while many at…
-
AI news: Robot flight delay, Doubao paid tiers, employee replaced by AI, and open-source controversy
A 35-year-old supervisor was replaced by AI at his financial technology company, leading to a salary reduction and subsequent termination. A court ruled in his favor, awarding him over 260,000 yuan in compensation, stat…
-
DeepClaude slashes coding agent costs by 17x using DeepSeek V4 Pro
An open-source tool called DeepClaude has gained significant traction by allowing developers to use the Claude Code agent loop with DeepSeek V4 Pro instead of Anthropic's models. This swap drastically reduces costs, wit…
-
LLMs Choose the Safer Gamble Yet Price the Riskier One Higher
A study involving four large language models—Claude Opus 4.7, DeepSeek V4-Pro, Google Gemini 3 Flash Preview, and OpenAI GPT-5.5—revealed a pattern of inconsistent decision-making. The models frequently chose a safer op…
-
Tech giants curb AI use as 'tokenmaxxing' drives up costs
Major tech companies like Microsoft, Meta, and Amazon are reportedly pulling back on internal AI usage due to escalating costs, primarily driven by the increased consumption of tokens by agentic AI tools. This phenomeno…
-
MoonPay 发布由 AI 驱动的 Mastercard，IBM 发布语音模型，DeepSeek V4 Pro 评估显示滞后
MoonPay 推出了 MoonAgents 卡，这是一种虚拟 Mastercard，允许 AI 代理使用来自自我托管钱包的稳定币进行直接支付。这一发展使 AI 系统能够在没有人工干预的情况下进行交易，实现财务自主。该卡支持与接受全球标准的任何商户进行支付，标志着 AI 金融自主堆栈的快速构建。
-
拥有 1.6T 参数的 DeepSeek-V4 Pro 模型现已上线 Together AI
DeepSeek-V4 Pro 是一个拥有 1.6 万亿参数的大型混合专家模型，现已在 Together AI 平台上可用。该模型专为长上下文推理而设计，在其最初的 Together AI 部署中支持高达 512K 令牌的上下文窗口，并计划支持 1M 令牌的上下文窗口。它具有可控的推理模式，可针对速度或深度进行优化，并为缓存的输入令牌提供专门的定价，以降低重复查询的成本。
-
小米的MiMo-v2.5-Pro开源模型可与顶级AI编码助手相媲美
小米发布了MiMo-v2.5-Pro，这是一款专注于编码的开源语言模型，在复杂任务中展现出令人印象深刻的能力。该模型在数小时内成功完成了一个大学级别的编译器项目，根据模糊的提示构建了一个功能齐全的视频编辑器应用程序，并解决了模拟电路设计问题。MiMo-v2.5-Pro在编码基准测试中表现强劲，可与GPT-5.4和Claude Opus 4.6等顶级闭源模型相媲美，现已在HuggingFace上发布。
-
Qwen 3.6 Plus outperforms DeepSeek V4 Pro in price and quality benchmarks
A recent battle test of six April-released Large Language Models (LLMs) revealed that the Qwen 3.6 Plus, released 22 days prior, outperformed the newer DeepSeek V4 Pro. Despite DeepSeek V4 Pro's advanced reasoning archi…
-
DeepSeek新AI模型市场反响平淡，竞争日益激烈
中国AI初创公司DeepSeek发布了其新DeepSeek-V4-Pro和DeepSeek-V4-Flash模型的预览版，但市场反应平淡。这与其之前推出的低成本AI模型（如DeepSeek-V3和DeepSeek-R1）获得的巨大关注形成鲜明对比，这些模型曾挑战了AI训练需要海量计算资源的必要性。分析人士认为，市场已经习惯了高效的模型开发，尽管V4模型有所改进，但与顶尖的开源竞争对手相比，其表现并未显著突出，尤其是在Kimi和Qwen…
-
Fireworks AI launches DeepSeek V4-Pro for inference infrastructure
DeepSeek V4-Pro, a new large language model, has been made available on the Fireworks AI inference platform. This release allows users to access and utilize the capabilities of the DeepSeek V4-Pro model through Firework…
-
Tencent's QClaw AI platform upgrades with Hermes support and new expert agents
Tencent's QClaw has released a significant update, version 0.2.14, introducing support for the Hermes Agent framework and integrating the DeepSeek-V4 Pro and Hy3 preview models. The platform has also enhanced its user e…
-
DeepSeek slashes API prices, making AI development cheaper
DeepSeek has reduced its API input cache hit cost by 90%, making development more affordable and efficient. This move coincides with an ongoing 75% discount promotion for their DeepSeek-V4-Pro model. Separately, David M…
-
Fireworks AI 在修复关键错误后发布 DeepSeek V4 Pro
Fireworks AI 发布了 DeepSeek V4 Pro，这是一个开源模型，在长上下文推理、代理性能和推理效率方面取得了显著进步。该模型采用混合专家架构和 1M token 上下文窗口，旨在以经济高效的方式处理广泛的状态和复杂的代理工作流。Fireworks AI 推迟了公开发布，以解决导致推理退化和输出损坏的关键服务路径正确性问题，确保在发布前已做好生产准备。