<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
        <title>TheRouter.ai — Новости AI</title>
        <link>https://therouter.ai/ru/news/</link>
        <description>AI-новости, релизы моделей и исследовательские материалы, собранные TheRouter.ai.</description>
        <lastBuildDate>Sat, 23 May 2026 00:00:00 GMT</lastBuildDate>
        <docs>https://validator.w3.org/feed/docs/rss2.html</docs>
        <generator>https://github.com/jpmonette/feed</generator>
        <language>ru-RU</language>
        <copyright>© 2026 TheRouter.ai</copyright>
        <atom:link href="https://therouter.ai/ru/news/rss.xml/" rel="self" type="application/rss+xml"/>
        <item>
            <title><![CDATA[DeepSeek Now Speaks Anthropic: What the New Dual-Format API Means for Your Routing Layer]]></title>
            <link>https://therouter.ai/ru/news/deepseek-anthropic-api-format-routing/</link>
            <guid>https://therouter.ai/ru/news/deepseek-anthropic-api-format-routing/</guid>
            <pubDate>Sat, 23 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[DeepSeek's API now accepts Anthropic SDK format at api.deepseek.com/anthropic — meaning Claude Code, the Anthropic Python/TS SDK, and any Anthropic-native client can now route requests to DeepSeek V4 models without an OpenAI wrapper.]]></description>
            <category>deepseek</category>
            <category>anthropic</category>
            <category>routing</category>
            <category>claude-code</category>
            <category>openai-compatible</category>
            <enclosure url="https://therouter.ai/news/deepseek-anthropic-api-format-routing/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[Anthropic Acquires Stainless: What SDK Consolidation Means for Multi-Provider API Teams]]></title>
            <link>https://therouter.ai/ru/news/anthropic-stainless-sdk-mcp-toolchain/</link>
            <guid>https://therouter.ai/ru/news/anthropic-stainless-sdk-mcp-toolchain/</guid>
            <pubDate>Fri, 22 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Anthropic has acquired Stainless, the company that generates every official Claude SDK and MCP server tooling. For teams building multi-provider API pipelines, this reshapes SDK dependency risk, MCP server governance, and the pace of Claude API surface changes.]]></description>
            <category>anthropic</category>
            <category>sdk</category>
            <category>mcp</category>
            <category>api</category>
            <category>developer-tooling</category>
            <enclosure url="https://therouter.ai/news/anthropic-stainless-sdk-mcp-toolchain/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[Kimi K2.6: Moonshot's Latest Open-Source Model Sets a New Bar for Long-Horizon Coding Agents]]></title>
            <link>https://therouter.ai/ru/news/kimi-k2-6-api-coding-agent-launch/</link>
            <guid>https://therouter.ai/ru/news/kimi-k2-6-api-coding-agent-launch/</guid>
            <pubDate>Fri, 22 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Moonshot AI releases Kimi K2.6 with state-of-the-art long-horizon coding, multimodal input (text, images, video), 256K context, and a fully OpenAI-compatible API — directly affecting how engineering teams route coding-agent workloads.]]></description>
            <category>kimi</category>
            <category>coding-agents</category>
            <category>openai-compatible</category>
            <category>model-routing</category>
            <category>provider-selection</category>
            <enclosure url="https://therouter.ai/news/kimi-k2-6-api-coding-agent-launch/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[Qwen-Image on DashScope: What the New Image Generation and Editing APIs Mean for Your Async Media Pipeline]]></title>
            <link>https://therouter.ai/ru/news/qwen-image-2-api-routing-async-media/</link>
            <guid>https://therouter.ai/ru/news/qwen-image-2-api-routing-async-media/</guid>
            <pubDate>Fri, 22 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Alibaba launched Qwen-Image and Qwen-Image-Edit on DashScope as qwen-image-2.0-pro. For routing teams, these introduce a non-DALL-E async job pattern and a new model namespace that standard OpenAI-compatible proxies may not handle correctly.]]></description>
            <category>qwen</category>
            <category>image-generation</category>
            <category>async-media</category>
            <category>routing</category>
            <category>multimodal</category>
            <enclosure url="https://therouter.ai/news/qwen-image-2-api-routing-async-media/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[Qwen-MT Turbo: Alibaba's Dedicated Translation API Introduces extra_body Routing Parameters That Standard Proxies May Drop]]></title>
            <link>https://therouter.ai/ru/news/qwen-mt-turbo-translation-api-routing/</link>
            <guid>https://therouter.ai/ru/news/qwen-mt-turbo-translation-api-routing/</guid>
            <pubDate>Fri, 22 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Alibaba Cloud's new Qwen-MT turbo model arrives via OpenAI-compatible endpoints, but its translation controls live inside extra_body — a pattern that breaks any middleware that strips non-standard fields. Here's what routing teams need to watch.]]></description>
            <category>qwen</category>
            <category>routing</category>
            <category>openai-compatible</category>
            <category>translation</category>
            <category>provider-ops</category>
            <enclosure url="https://therouter.ai/news/qwen-mt-turbo-translation-api-routing/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[Qwen3.7-Max Launches with Top Agent Benchmarks: What Routing Teams Need to Know]]></title>
            <link>https://therouter.ai/ru/news/qwen3-7-max-agent-benchmark-launch/</link>
            <guid>https://therouter.ai/ru/news/qwen3-7-max-agent-benchmark-launch/</guid>
            <pubDate>Fri, 22 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Alibaba Cloud's Qwen3.7-Max claims the top spot on Terminal Bench 2.0, outperforms Claude Opus 4.6 on GPQA Diamond, and completes a 35-hour autonomous agent task on novel hardware. Here's the operator angle for teams routing Qwen traffic.]]></description>
            <category>qwen</category>
            <category>agent</category>
            <category>provider-ops</category>
            <category>openai-compatible</category>
            <enclosure url="https://therouter.ai/news/qwen3-7-max-agent-benchmark-launch/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[DeepSeek's Official Coding Agent Guide: Route Claude Code and OpenCode to V4 Models]]></title>
            <link>https://therouter.ai/ru/news/deepseek-coding-agent-integration-guide-claude-code-opencode/</link>
            <guid>https://therouter.ai/ru/news/deepseek-coding-agent-integration-guide-claude-code-opencode/</guid>
            <pubDate>Thu, 21 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[DeepSeek published an official integration guide for Claude Code, OpenCode, and OpenClaw — revealing a per-tier model routing pattern that operators can apply across any Anthropic-compatible gateway.]]></description>
            <category>deepseek</category>
            <category>claude-code</category>
            <category>coding-agents</category>
            <category>routing</category>
            <category>openai-compatible</category>
            <enclosure url="https://therouter.ai/news/deepseek-coding-agent-integration-guide-claude-code-opencode/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[DeepSeek Retires deepseek-chat and deepseek-reasoner: What the V4 Model Rename Means for Routing Teams]]></title>
            <link>https://therouter.ai/ru/news/deepseek-v4-api-migration-deprecation-2026/</link>
            <guid>https://therouter.ai/ru/news/deepseek-v4-api-migration-deprecation-2026/</guid>
            <pubDate>Thu, 21 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[DeepSeek deprecates deepseek-chat and deepseek-reasoner on July 24, 2026. Routing configs must migrate to deepseek-v4-flash or deepseek-v4-pro. A new Anthropic-compatible endpoint also changes how Claude Code users wire DeepSeek into their agent workflows.]]></description>
            <category>deepseek</category>
            <category>api-deprecation</category>
            <category>routing</category>
            <category>claude-code</category>
            <enclosure url="https://therouter.ai/news/deepseek-v4-api-migration-deprecation-2026/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[Qwen3Guard: Alibaba's Open-Source Streaming Safety Guardrail for Multi-Provider AI Pipelines]]></title>
            <link>https://therouter.ai/ru/news/qwen3guard-streaming-safety-guardrail-open-source/</link>
            <guid>https://therouter.ai/ru/news/qwen3guard-streaming-safety-guardrail-open-source/</guid>
            <pubDate>Thu, 21 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Alibaba releases Qwen3Guard, an open-weight safety guardrail model with real-time streaming token-level detection. For teams routing across multiple AI providers, it offers a provider-agnostic safety layer you can deploy at the gateway level.]]></description>
            <category>governance</category>
            <category>qwen</category>
            <category>open-source</category>
            <category>routing</category>
            <enclosure url="https://therouter.ai/news/qwen3guard-streaming-safety-guardrail-open-source/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[AI Cost Increases 2026: Why List Price is No Longer Enough]]></title>
            <link>https://therouter.ai/ru/news/ai-cost-increases-claude-opus-gpt-5-5-2026/</link>
            <guid>https://therouter.ai/ru/news/ai-cost-increases-claude-opus-gpt-5-5-2026/</guid>
            <pubDate>Tue, 19 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[OpenAI, Anthropic, and GitHub changed pricing models in the same week. List price gaps reach 92% depending on tokenizer behavior and usage patterns. Routing architectures are now essential for cost control.]]></description>
            <category>pricing</category>
            <category>routing</category>
            <category>openai</category>
            <category>anthropic</category>
            <enclosure url="https://therouter.ai/news/ai-cost-increases-claude-opus-gpt-5-5-2026/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[OpenAI Embeds C2PA + SynthID in Every API-Generated Image: What It Means for Your AI Pipeline]]></title>
            <link>https://therouter.ai/ru/news/openai-c2pa-synthid-api-image-provenance/</link>
            <guid>https://therouter.ai/ru/news/openai-c2pa-synthid-api-image-provenance/</guid>
            <pubDate>Tue, 19 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[OpenAI is now a C2PA conforming generator with Google SynthID watermarks in every API-generated image. Every output carries a cryptographic provenance fingerprint — affecting audit trails, fallback routing, and enterprise content governance.]]></description>
            <category>openai</category>
            <category>api</category>
            <category>governance</category>
            <category>content-provenance</category>
            <category>enterprise</category>
            <category>security</category>
            <enclosure url="https://therouter.ai/news/openai-c2pa-synthid-api-image-provenance/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[Qwen GSPO: Sequence-Level RL Stabilizes Large-Scale Language Model Training]]></title>
            <link>https://therouter.ai/ru/news/qwen-gspo-sequence-level-rllm-stability/</link>
            <guid>https://therouter.ai/ru/news/qwen-gspo-sequence-level-rllm-stability/</guid>
            <pubDate>Tue, 19 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[GSPO introduces sequence-level optimization for RL training, stabilizing MoE models and eliminating infrastructure-heavy Routing Replay workarounds.]]></description>
            <category>reinforcement-learning</category>
            <category>training-infrastructure</category>
            <category>qwen</category>
            <category>moe</category>
            <enclosure url="https://therouter.ai/news/qwen-gspo-sequence-level-rllm-stability/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[US AI Model Oversight Would Turn Release Timing Into a Routing Risk]]></title>
            <link>https://therouter.ai/ru/news/us-ai-model-oversight-routing-risk/</link>
            <guid>https://therouter.ai/ru/news/us-ai-model-oversight-routing-risk/</guid>
            <pubDate>Mon, 18 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[A reported US review process for frontier AI models would make provider availability, release timing, and governance evidence part of routing strategy—not just policy news.]]></description>
            <category>ai-governance</category>
            <category>routing</category>
            <category>provider-risk</category>
            <enclosure url="https://therouter.ai/news/us-ai-model-oversight-routing-risk/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[AI-Generated Audemars Piguet Royal Oak Images Drive Week-Long Hype Before Real Royal Pop Launch]]></title>
            <link>https://therouter.ai/ru/news/ai-generated-audemars-piguet-royal-oak-images-drive-week-long-hype-before-real-royal-pop-launch/</link>
            <guid>https://therouter.ai/ru/news/ai-generated-audemars-piguet-royal-oak-images-drive-week-long-hype-before-real-royal-pop-launch/</guid>
            <pubDate>Mon, 18 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[AI-generated images of colorful Audemars Piguet Royal Oak wristwatches flooded social media for a week before Swatch's actual Royal Pop pocket watch collaboration was revealed, creating unprecedented fake product hype.]]></description>
            <enclosure url="https://therouter.ai/news/ai-generated-audemars-piguet-royal-oak-images-drive-week-long-hype-before-real-royal-pop-launch/cover.jpg/" length="0" type="image/jpg/"/>
        </item>
        <item>
            <title><![CDATA[OpenAI Musk Altman Trial Features Donkey Statue as Evidence in Safety Dispute]]></title>
            <link>https://therouter.ai/ru/news/openai-musk-altman-trial-features-donkey-statue-as-evidence-in-safety-dispute/</link>
            <guid>https://therouter.ai/ru/news/openai-musk-altman-trial-features-donkey-statue-as-evidence-in-safety-dispute/</guid>
            <pubDate>Mon, 18 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[OpenAI sought to present a jackass trophy in the Musk v. Altman trial as evidence of Elon Musk's behavior toward employees who challenged his safety priorities in 2018.]]></description>
            <enclosure url="https://therouter.ai/news/openai-musk-altman-trial-features-donkey-statue-as-evidence-in-safety-dispute/cover.jpg/" length="0" type="image/jpg/"/>
        </item>
        <item>
            <title><![CDATA[Sam Altman Testifies Elon Musk Had 'Hair-Raising' Idea to Pass OpenAI to His Kids]]></title>
            <link>https://therouter.ai/ru/news/sam-altman-testifies-elon-musk-had-hair-raising-idea-to-pass-openai-to-his-kids/</link>
            <guid>https://therouter.ai/ru/news/sam-altman-testifies-elon-musk-had-hair-raising-idea-to-pass-openai-to-his-kids/</guid>
            <pubDate>Mon, 18 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[OpenAI CEO Sam Altman testified in the Musk v. Altman trial, revealing Elon Musk's suggestion to pass OpenAI control to his children and defending against allegations of deceptive behavior.]]></description>
            <enclosure url="https://therouter.ai/news/sam-altman-testifies-elon-musk-had-hair-raising-idea-to-pass-openai-to-his-kids/cover.jpg/" length="0" type="image/jpg/"/>
        </item>
        <item>
            <title><![CDATA[WhatsApp Incognito Chat Launches with Meta AI Private Processing]]></title>
            <link>https://therouter.ai/ru/news/whats-app-incognito-chat-launches-with-meta-ai-private-processing/</link>
            <guid>https://therouter.ai/ru/news/whats-app-incognito-chat-launches-with-meta-ai-private-processing/</guid>
            <pubDate>Mon, 18 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[WhatsApp introduces Incognito Chat for Meta AI conversations, using Private Processing to prevent Meta from accessing user interactions with the chatbot.]]></description>
            <enclosure url="https://therouter.ai/news/whats-app-incognito-chat-launches-with-meta-ai-private-processing/cover.jpg/" length="0" type="image/jpg/"/>
        </item>
        <item>
            <title><![CDATA[Musk v. Altman Trial Closing Arguments Expose OpenAI Nonprofit Mission Tensions]]></title>
            <link>https://therouter.ai/ru/news/musk-v-altman-trial-closing-arguments-expose-openai-nonprofit-mission-tensions/</link>
            <guid>https://therouter.ai/ru/news/musk-v-altman-trial-closing-arguments-expose-openai-nonprofit-mission-tensions/</guid>
            <pubDate>Fri, 15 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[The Musk v. Altman trial closing arguments reveal how OpenAI's nonprofit structure conflicted with competitive pressures, raising questions about public interest protection.]]></description>
            <enclosure url="https://therouter.ai/news/musk-v-altman-trial-closing-arguments-expose-openai-nonprofit-mission-tensions/cover.jpg/" length="0" type="image/jpg/"/>
        </item>
        <item>
            <title><![CDATA[AI Sustainability Research Pushes for Transparency on Model Energy Use]]></title>
            <link>https://therouter.ai/ru/news/ai-sustainability-research-pushes-for-transparency-on-model-energy-use/</link>
            <guid>https://therouter.ai/ru/news/ai-sustainability-research-pushes-for-transparency-on-model-energy-use/</guid>
            <pubDate>Wed, 13 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Sasha Luccioni launches Sustainable AI Group to address growing enterprise and regulatory demands for AI emissions data and energy efficiency metrics across European and global markets.]]></description>
            <enclosure url="https://therouter.ai/news/ai-sustainability-research-pushes-for-transparency-on-model-energy-use/cover.jpg/" length="0" type="image/jpg/"/>
        </item>
        <item>
            <title><![CDATA[OpenAI's TanStack Supply Chain Breach: What the Codex CLI Cert Rotation Means for Your Dev Toolchain]]></title>
            <link>https://therouter.ai/ru/news/openai-tanstack-supply-chain-codex-cli-security/</link>
            <guid>https://therouter.ai/ru/news/openai-tanstack-supply-chain-codex-cli-security/</guid>
            <pubDate>Wed, 13 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[OpenAI confirmed two employee devices were hit by the Mini Shai-Hulud npm supply chain attack, exposing code-signing certificates for Codex CLI, ChatGPT Desktop, and Atlas. macOS users must update before June 12 or their apps will stop working.]]></description>
            <category>security</category>
            <category>openai</category>
            <category>codex</category>
            <category>supply-chain</category>
            <category>developer-tools</category>
            <enclosure url="https://therouter.ai/news/openai-tanstack-supply-chain-codex-cli-security/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[Ilya Sutskever Testifies in Musk v Altman Trial: 'I Didn't Want OpenAI to Be Destroyed']]></title>
            <link>https://therouter.ai/ru/news/ilya-sutskever-testifies-in-musk-v-altman-trial-i-didnt-want-openai-to-be-destroyed/</link>
            <guid>https://therouter.ai/ru/news/ilya-sutskever-testifies-in-musk-v-altman-trial-i-didnt-want-openai-to-be-destroyed/</guid>
            <pubDate>Tue, 12 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Former OpenAI chief scientist Ilya Sutskever defended his role in Sam Altman's ouster during testimony in the Musk v Altman trial, stating he acted to protect the company from destruction.]]></description>
            <enclosure url="https://therouter.ai/news/ilya-sutskever-testifies-in-musk-v-altman-trial-i-didnt-want-openai-to-be-destroyed/cover.webp/" length="0" type="image/webp/"/>
        </item>
        <item>
            <title><![CDATA[TheRouter v2.0: Unified Async Media Platform for Image, Audio, and Video]]></title>
            <link>https://therouter.ai/ru/news/therouter-async-media-platform-v2/</link>
            <guid>https://therouter.ai/ru/news/therouter-async-media-platform-v2/</guid>
            <pubDate>Fri, 08 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[TheRouter v2.0 introduces a unified async paradigm across image, audio, and video generation. Long-running upstream calls move off the request thread, into a job-based flow with S3-backed artifacts and a single /v1/jobs/:id polling surface.]]></description>
            <category>therouter</category>
            <category>async</category>
            <category>media</category>
            <category>release</category>
            <enclosure url="https://therouter.ai/news/therouter-async-media-platform-v2/cover.png/" length="0" type="image/png/"/>
        </item>
        <item>
            <title><![CDATA[AI-Coded Apps from Lovable, Replit, Base44 Expose Corporate Data on Open Web]]></title>
            <link>https://therouter.ai/ru/news/ai-coded-apps-from-lovable-replit-base44-expose-corporate-data-on-open-web/</link>
            <guid>https://therouter.ai/ru/news/ai-coded-apps-from-lovable-replit-base44-expose-corporate-data-on-open-web/</guid>
            <pubDate>Thu, 07 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Security researchers discovered over 5,000 AI-coded apps from Lovable, Replit, Base44, and Netlify exposing sensitive corporate and personal data without authentication or security measures.]]></description>
            <enclosure url="https://therouter.ai/news/ai-coded-apps-from-lovable-replit-base44-expose-corporate-data-on-open-web/cover.jpg/" length="0" type="image/jpg/"/>
        </item>
        <item>
            <title><![CDATA[Anthropic Launches Claude Managed Agents Platform for Enterprise AI]]></title>
            <link>https://therouter.ai/ru/news/anthropic-managed-agents-platform/</link>
            <guid>https://therouter.ai/ru/news/anthropic-managed-agents-platform/</guid>
            <pubDate>Thu, 07 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Anthropic introduces a managed runtime for Claude agents, handling tool use, memory, and orchestration so enterprise teams can ship agents without managing the runtime themselves.]]></description>
            <category>anthropic</category>
            <category>agents</category>
            <category>enterprise</category>
            <enclosure url="https://therouter.ai/news/anthropic-managed-agents-platform/cover.png/" length="0" type="image/png/"/>
        </item>
        <item>
            <title><![CDATA[Meta Ray-Ban Smart Glasses Face Recognition Feature Opposed by 70+ Civil Rights Groups]]></title>
            <link>https://therouter.ai/ru/news/meta-ray-ban-smart-glasses-face-recognition-feature-opposed-by-70-civil-rights-groups/</link>
            <guid>https://therouter.ai/ru/news/meta-ray-ban-smart-glasses-face-recognition-feature-opposed-by-70-civil-rights-groups/</guid>
            <pubDate>Mon, 13 Apr 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[More than 70 civil liberties organizations demand Meta abandon facial recognition plans for Ray-Ban smart glasses, warning the 'Name Tag' feature would enable stalkers and predators to identify strangers in public.]]></description>
            <enclosure url="https://therouter.ai/news/meta-ray-ban-smart-glasses-face-recognition-feature-opposed-by-70-civil-rights-groups/cover.jpg/" length="0" type="image/jpg/"/>
        </item>
    </channel>
</rss>