149 public catches this week · all weeks · the live wire

10-04URGENTbinary

New strings in the Claude code binary suggest updates to plugin management, UI notifications, and error handling, alongside a new model ID reference and endpoints.

10-03RTNgithub_releases+pages

Anthropic's claude-code released v2.1.289 with numerous bug fixes related to shell commands, terminal rendering, plugin behavior, and error handling, while also adding `agent.spawn` for teammates.

10-03URGENTbinary

Anthropic's Claude-Code binary update reveals a new `claude-sonnet-4` model ID, several internal feature flags related to multi-turn interactions, plugin UI, and session limits, and new API endpoints.

10-02RTNdocs

Claude Code updated to 2.1.288, adding new UI features like selection retrieval for mods and session search, alongside numerous bug fixes for prompt handling, session resumption, and plugin stability.

10-02URGENTgithub_releases+npm

Claude Code v2.1.288 release includes a new `$.ui.selection()` mod for selected text, built-in `gh api` for cloud sessions, and various bug fixes for prompt recovery, API timeouts, and conversation length.

10-02FLASHlmarena

LMArena has listed four new models, including 'claude-sonnet-5-5-max-agent' and 'step-5-preview-agent', potentially indicating upcoming releases or tests. ✓ beat the announcement by 0.0h

10-02RTNweb app

Anthropic's Claude.com frontend deployed new code including a Stripe API endpoint for payment intents, suggesting upcoming payment processing functionality.

10-01URGENTsitemap

Anthropic's sitemap shows 20 new beta API pages, indicating a potential analytics and role-based access control (RBAC) expansion.

10-01RTNweb app

Anthropic updated their Claude console frontend, adding references to hCaptcha, potentially for enhanced security or anti-bot measures.

10-01FLASHmodel_catalog

13 new model IDs detected across multiple providers including perplexity, nebius, openrouter, and voyage, hinting at unannounced models like GPT-6 variations, Grok-4.7, and new Claude versions.

10-01URGENTbinary

Anthropic's native binary for Claude Code on macOS ARM64 was updated, adding references to a new 'claude-sonnet-5-5' model with 1M context, new API endpoints, and several internal feature flags related to worker interactions and cloud sessions. ✓ beat the announcement by 22.6h

10-01FLASHgithub_releases+pages

Claude Code released version 2.1.287, introducing "Claude Mods" for deeper plugin behavior modification and a new "You should know" plugin for a side agent to flag potential misses.

10-01URGENTsitemap

Anthropic added 12 new, unlinked sitemap URLs for 'plugins-api' documentation in multiple languages, suggesting upcoming plugin functionality for Claude.

10-01FLASHlmarena

LMArena has added a new model, 'claude-sonnet-5-5-xhigh', to its leaderboard, suggesting an upcoming release from Anthropic. ✓ beat the announcement by 25.0h

10-01URGENThacker news

FTC is investigating OpenAI, Anthropic, and other AI companies regarding product risks, as indicated by a fast-rising Hacker News story.

10-01URGENTbinary

Anthropic's `claude-code` binary adds strings referencing "claude-haiku-3-5" and a "feature_multi_turn_ask" flag, suggesting new model and multi-turn interaction features.

09-30FLASHsitemap

Anthropic quietly published 98 new documentation pages related to API organization, plugin marketplaces, and plugins, suggesting a significant expansion in their platform's extensibility and management features.

09-30RTNgithub

Anthropic's Python SDK v1.11.0 adds a 'list spend limits' endpoint to the API and deprecates 'sonnet 4.5'.

09-30FLASHpypi

Anthropic's 1.11.0 client update includes references to new model IDs "claude-opus-4-8" and "claude-haiku-4-5", along with new beta features for messages streaming and environments.

09-30URGENTnpm

Anthropic's SDK update includes new string literals related to tool use, chat examples, and a `myContext` request option.

09-30URGENTgithub

Anthropic's Python SDK v1.10.0 includes new Admin API features for Claude Enterprise analytics, spend limits, RBAC, plugins, and per-user usage reports.

09-30RTNgithub_releases+pages

Anthropic's Claude Code v2.1.286 release includes fixes for login issues, tool call handling, session recovery, and security enhancements for logs and transcripts.

09-30FLASHpypi

Anthropic's 1.10.0 release includes new model IDs for 'claude-opus-4-8' and 'claude-haiku-4-5', along with numerous beta endpoints for analytics, plugins, and RBAC.

09-30URGENTnpm

Anthropic's SDK update 0.130.0 includes numerous beta endpoints related to organizational analytics, plugin marketplaces, RBAC, and spend limits, alongside a new beta feature flag for 'ce-plugins-2026-09-01'.

09-30URGENTbinary

Anthropic's Claude Code for macOS arm64 (2.1.285) binary includes a new model ID and numerous beta feature flags related to organizations, rate limits, and service accounts, suggesting upcoming API and UI features.

09-29FLASHgithub_releases+npm+pages

Anthropic's Claude Code v2.1.285 adds new environment variables, plugin configuration options, and fixes numerous issues, including subagent and plugin management.

09-29FLASHmodel_catalog

Four new model IDs, including `global.openai.gpt-6.1-sol`, `gpt-4o-mini-tts-2025-03-20`, `openai.gpt-6.1-sol`, and `openrouter/anthropic/claude-sonnet-5.5:batch`, appeared in the Litellm model catalog. ✓ beat the announcement by 21.0h

09-29FLASHlmarena

A new Claude model, claude-sonnet-5-5-high, appeared on the LMArena leaderboard, suggesting a potential upcoming release. ✓ beat the announcement by 39.2h

09-29URGENThacker news

A Hacker News story about a partial Claude outage is rapidly gaining traction, indicating potential service disruptions.

09-29FLASHmodel_catalog

7 new model IDs were added to the LiteLLM catalog, including new Claude Opus and Sonnet versions on Bedrock Mantle, and new Fireworks AI router models.

09-29RTNweb app

Anthropic's Claude platform deployed new frontend code, including a new endpoint related to OAuth authorization redirects.

09-29RTNhacker news

Anthropic's IPO prospectus hints at AI vision capabilities and increasing operational costs, gaining rapid traction on Hacker News.

09-28FLASHsitemap

Anthropic quietly added 13 new pages to its docs sitemap, hinting at a 'Claude Sonnet 5' system prompt and an 'Access Transparency Log' feature.

09-28FLASHapi_models+model_catalog

OpenRouter's API now lists 'anthropic/claude-sonnet-5.5:batch', suggesting a new unannounced Claude Sonnet batch processing model.

09-28URGENTbinary

Anthropic's `claude-code-darwin-arm64` binary update hints at new `claude-sonnet-5-5` and `claude-opus-4-5-20251101` models, alongside several new internal feature flags and endpoints related to usage and security.

09-28RTNnpm

Anthropic's claude-code package updated with new strings referencing artifact URLs, suggesting a potential feature for in-place updates of artifacts via claude.ai links.

09-28FLASHgithub_releases+hn+model_catalog+npm+pages+pypi+sitemap

Anthropic's SDK update 0.129.0 suggests new `claude-sonnet-5-5` model with tool runner compaction, and new `POST /v1/messages` and `POST /v1/messages/count_tokens` endpoints. ✓ beat the announcement by 0.0h

09-28FLASHmodel api

OpenRouter API shows a new model ID: anthropic/claude-sonnet-5.5, suggesting an upcoming release from Anthropic. ✓ beat the announcement by 0.0h

10-04RTNnpm

OpenAI's latest SDK update (7.28.0) hints at new file search tool configurations including chunking options and a 'ranker' parameter. It also emphasizes JSON mode usage.

10-04RTNgithub

OpenAI Codex released a new pre-release version for its Rust SDK, v0.162.0-alpha.13, indicating ongoing development.

10-04RTNgithub

OpenAI Codex released a new pre-release version: rust-v0.162.0-alpha.12.

10-04URGENThacker news

An OpenAI safety leader has reportedly quit, citing a "broken" company culture, as indicated by a rapidly climbing Hacker News story.

10-03RTNgithub

OpenAI's Codex project released a new pre-release version: rust-v0.162.0-alpha.11, indicating ongoing development, possibly in Rust.

10-03FLASHmodel_catalog

Six new model IDs detected, including several new 'gpt-image' models via OpenRouter/OpenAI, a 'grok-imagine-video' model from XAI, and Cohere's new embed-v5.0-fast and pro.

10-03URGENThacker news

An Atlantic article titled "I Quit OpenAI Because Its Culture Is Broken" is gaining significant traction on Hacker News, suggesting potential internal issues at OpenAI.

10-03RTNgithub

OpenAI's Codex project released a new pre-release version: rust-v0.162.0-alpha.10, indicating continued development on its Rust integration.

10-03URGENTgithub

A new alpha pre-release, rust-v0.162.0-alpha.9, has been published for openai/codex. ✓ beat the announcement by 26.9h

10-02RTNgithub

OpenAI Codex published a new pre-release for its Rust SDK: rust-v0.162.0-alpha.8.

10-02RTNgithub

OpenAI's Python SDK v3.24.0 adds support for custom voice creation and agent session events, alongside minor bug fixes for request routing.

10-02FLASHpypi

OpenAI's latest SDK update (3.24.0) hints at a new model, 'gpt-4o-2024-08-06', alongside new 'realtime' and 'responses' beta API connections, and expanded agent session event types.

10-02FLASHlmarena

LMArena has added four new models to its leaderboard, including one anonymous model (`fiddler-l4ag`) and three named models (`contenders/gpt-6.1-sol-max-agent`, `gpt-6.1-sol-max`, and `gpt-6.1-sol-max-agent`), with the anonymous model potentially being a stealth release. ✓ beat the announcement by 0.0h

10-02RTNgithub

OpenAI's Codex project has released a new alpha version, rust-v0.162.0-alpha.7, indicating ongoing development in its Rust implementation.

10-02URGENTgithub

OpenAI's Codex project published a new pre-release, rust-v0.162.0-alpha.6. ✓ beat the announcement by 39.9h

10-02URGENTgithub

OpenAI Codex published a pre-release for its Rust library, version 0.162.0-alpha.5. ✓ beat the announcement by 44.0h

10-02RTNgithub

OpenAI's Codex project released a new pre-release version: rust-v0.162.0-alpha.4.

10-02URGENTgithub

OpenAI's Codex project published a new pre-release: rust-v0.162.0-alpha.3, suggesting ongoing development in Rust. ✓ beat the announcement by 48.0h

10-02RTNgithub

OpenAI's Codex project published a new pre-release: rust-v0.162.0-alpha.2. This is an alpha version, indicating early development or testing.

10-02RTNweb app

A new endpoint `https://chat.openai.com/` was added to the ChatGPT web app frontend bundle.

10-01RTNgithub

OpenAI's Codex project released a new pre-release version, rust-v0.162.0-alpha.1, indicating ongoing development in a Rust-based component.

10-01URGENTgithub

OpenAI's python library (v3.23.0) adds new agent capabilities including returning typed answers from session streams, staging files, and downloading turn artifacts, as well as session traces and Realtime translations.

10-01FLASHpypi

OpenAI's latest SDK update (3.23.0) introduces a new model ID, gpt-4o-2024-08-06, and significant beta features around an Agents API for hosted tools, real-time connections, and artifact management.

10-01URGENTnpm

OpenAI's new npm release includes strings for a new `ranker` option with a `default_2024_08_21` value, suggesting an updated ranking algorithm for file search. It also hints at new `code_interpreter` and `file_search` tool types with detailed configurations for chunking and overlap tokens.

10-01URGENTnpm

OpenAI's SDK update (v7.26.0) hints at new `file_search` and `code_interpreter` tool types, a 'ranker' option, and JSON mode for models, along with new event streaming capabilities.

10-01URGENTgithub

OpenAI's Codex project published a new pre-release for its Rust-based version, 0.159.0-alpha.12.1. ✓ beat the announcement by 56.1h

10-01RTNgithub

OpenAI Codex added opt-in Guardian review capabilities for retrieving user instructions and agent handoff context, alongside other minor features and bug fixes in rust-v0.160.0.

10-01RTNgithub

OpenAI Codex published a pre-release for rust-v0.161.0-alpha.13, indicating ongoing development for their code generation model.

10-01URGENTgithub

OpenAI's Codex project published a new pre-release: rust-v0.161.0-alpha.12. ✓ beat the announcement by 1.0h

10-01URGENTgithub

OpenAI's Codex project published a new pre-release: rust-v0.161.0-alpha.11. ✓ beat the announcement by 3.8h

10-01RTNgithub

OpenAI's Codex repository published a new pre-release for its Rust client, version 0.161.0-alpha.10.

10-01URGENTgithub

OpenAI Codex released a new pre-release version for its Rust integration: `rust-v0.161.0-alpha.9`. ✓ beat the announcement by 7.0h

10-01RTNgithub

OpenAI's Codex project published a new alpha pre-release for its Rust-based version, v0.161.0-alpha.8.

10-01RTNgithub

OpenAI's Codex project published a new pre-release version: rust-v0.161.0-alpha.7.

10-01URGENTgithub

OpenAI's Codex project released a new pre-release version: rust-v0.161.0-alpha.6. ✓ beat the announcement by 13.8h

10-01RTNgithub

OpenAI Codex published a pre-release for its Rust library, version 0.160.0-alpha.6.2.

09-30RTNgithub

OpenAI's codex rust library release 0.159.3 adds optional security setup reminders for ChatGPT local sessions.

09-30RTNgithub

OpenAI's Codex project has published a new pre-release for its Rust-based version, `rust-v0.161.0-alpha.5`.

09-30FLASHlmarena

A new model, gpt-6.1-sol-max-code-codex-harness, appeared on the LMArena leaderboard, suggesting a potential unreleased GPT-6.1 variant focused on code. ✓ beat the announcement by 0.9h

09-30RTNgithub

OpenAI's Codex project released a new pre-release version: rust-v0.161.0-alpha.4.

09-30RTNgithub

A new alpha pre-release, rust-v0.161.0-alpha.3, has been published for openai/codex.

09-30RTNgithub

OpenAI Codex published a new pre-release, rust-v0.160.0-alpha.6.1, indicating continued development in Rust for Codex.

09-30FLASHpypi

OpenAI's 3.22.1 update hints at a new model, gpt-4o-2024-08-06, alongside new 'realtime' and 'responses' beta features, and expanded streaming capabilities for chat completions and threads.

09-29FLASHmodel_catalog

Two new OpenAI GPT-6.1-sol model IDs (bedrock_mantle/openai.gpt-6.1-sol and us.openai.gpt-6.1-sol) have appeared in the LiteLLM model catalog, suggesting potential upcoming releases or integrations across different cloud providers. ✓ beat the announcement by 18.0h

09-29FLASHgithub

OpenAI's Codex project has released rust-v0.159.1, featuring the addition of GPT-6.1 Sol as the new default model in its bundled, Amazon Bedrock Mantle, and Bedrock Runtime catalogs. ✓ beat the announcement by 20.0h

09-29URGENTgithub

OpenAI's Codex project has published a new alpha pre-release for its Rust-based version, `rust-v0.161.0-alpha.2`.

09-29RTNgithub

OpenAI's Python library (v3.22.0) adds 'computer use' feature to beta agents.

09-29FLASHpypi

OpenAI's latest SDK includes a new model ID 'gpt-4o-2024-08-06' and new beta real-time and responses connection features.

09-29URGENTnpm

OpenAI's SDK update 7.25.0 includes new file search tool overrides with chunking and static options, a new 'default_2024_08_21' ranker, and hints at event emissions for object lifecycle and streaming deltas.

09-29URGENThacker news

An OpenAI help article titled "About ChatGPT Pro Tiers" is generating significant interest on Hacker News, suggesting potential new paid tiers for ChatGPT.

09-29URGENTgithub_org

OpenAI released a new public repository: 'sign-in-with-chatgpt-devkit', which includes an SDK, React components, and design assets for integrating 'Sign-in with ChatGPT' functionality.

09-29FLASHgithub

OpenAI's Python library adds a new model identifier, GPT-6.1 Sol, in release v3.21.0. ✓ beat the announcement by 0.0h

09-29FLASHapi_models+model_catalog+npm+pages+pypi

OpenAI's latest npm package (7.24.0) hints at new `gpt-6.1-sol` models and introduces `file_search` tool overrides with chunking controls, along with new realtime translation WebSocket features. ✓ beat the announcement by 0.0h

09-29FLASHhacker news

An OpenAI announcement titled "Dots" is rapidly gaining traction on Hacker News, suggesting a significant new release or product. ✓ beat the announcement by 0.0h

09-29FLASHhn+sitemap

OpenAI's sitemap reveals 8 new pages, including "introducing gpt-6-1-sol" and new forms, corroborated by a fast-climbing Hacker News story about GPT-6.1 Sol. ✓ beat the announcement by 0.0h

09-29FLASHdocs

OpenAI's changelog received a significant update, hinting at new features like 'Agents', 'Dots', 'Fast mode'/'Ultrafast mode', 'Computer use', 'Bedrock Managed Agents', and an expanded focus on ChatGPT integration and developer tools. ✓ beat the announcement by 1.1h

09-29URGENTgithub_org

OpenAI released a new public repository 'mcp-extensions' to build ChatGPT plugins that integrate like native features.

09-29RTNgithub

OpenAI Codex published a pre-release for its Rust library, version 0.161.0-alpha.1.

09-29RTNgithub

OpenAI's Codex rust-v0.159.0 release adds opt-in `instant_interrupt` for steering during model responses, improves welcome screens and warnings, enhances transcript scrolling, and expands Mermaid rendering.

09-29RTNgithub

OpenAI's Codex project has released a new pre-release version: rust-v0.160.0-alpha.6, indicating ongoing development in its Rust implementation.

09-29URGENTgithub

OpenAI's Codex project has released a new Rust pre-release version: rust-v0.160.0-alpha.3. ✓ beat the announcement by 5.0h

09-28URGENTgithub

OpenAI's Codex project released a new pre-release version: rust-v0.160.0-alpha.2. ✓ beat the announcement by 10.3h

09-28RTNhacker news

An HN story is rapidly gaining traction, suggesting OpenAI is struggling to manage 'rogue AI activity.'

09-28FLASHlmarena

LMArena's leaderboard shows two new models: 'contenders/gpt-6-luna-max-agent' and 'gpt-6-luna-max-agent', possibly indicating unreleased model testing. ✓ beat the announcement by 0.0h

09-28URGENTgithub_org

OpenAI made a new public repository called 'openai-live-console' which appears to be a local voice console with read-only tools and backend delegation.

09-28RTNgithub

OpenAI Python SDK v3.20.0 adds features for Agents credentials, Cyber access programs, and incremental WebSocket text/tool snapshots, alongside various bug fixes and documentation updates.

09-28FLASHpypi

OpenAI's latest SDK includes a new 'gpt-4o-2024-08-06' model ID, and adds new beta `realtime` and `responses` connection features, hinting at enhanced streaming or interactive capabilities.

09-28RTNgithub

OpenAI's Codex project published a new pre-release version: rust-v0.159.0-alpha.13, indicating ongoing development, possibly for a Rust-based component or API.

09-28RTNweb app

ChatGPT web app is referencing a new 'chatgpt-staging.com' endpoint, suggesting internal staging environment setup.

09-28RTNgithub

OpenAI Codex's `rust-v0.158.0` release adds explicit transparent backgrounds for image generation, secure WebSocket connections with bearer tokens, and improves TUI copy-on-select functionality.

09-28FLASHweb app

ChatGPT web app updated with potential references to new GPT-4.1, GPT-5, GPT-5.1, GPT-5.2, GPT-5.3, GPT-5.4, GPT-5.5, GPT-5.6, and GPT-4o-gdrive models, alongside internal feature flags and endpoints. ✓ beat the announcement by 13.1h

09-28RTNgithub

OpenAI Codex released a new pre-release version for its Rust integration, `rust-v0.159.0-alpha.12`.

09-28URGENTgithub

OpenAI's Codex project published a new pre-release version: rust-v0.158.0-alpha.15.4. ✓ beat the announcement by 28.0h

09-28RTNgithub

OpenAI's Codex project published a new pre-release for its Rust SDK, version 0.159.0-alpha.11.

10-04FLASHhuggingface

Google pushed a new HuggingFace repository for a model called DiarizationLM-Gemma-4-E4B-v1, likely indicating a new language model release related to diarization.

10-02RTNgithub

Gemini CLI released a new nightly version v0.64.0-nightly.20261002.gc9096a847 with fixes for chat recording, atomic state persistence, Ctrl+C handling, and UI issues.

10-01FLASHmodel_catalog

A new model ID, grok-4.7, appeared in the LiteLLM model catalog under Google Cloud's Vertex AI/XAI.

10-01RTNgithub

Gemini CLI pre-release v0.64.0-nightly.20261001.gc6bccb7ec fixes CPU hang with `@` in code and makes file tool writes atomic.

09-30FLASHhacker news

A new model named "Gemini 4 Argon" is trending on Hacker News, indicating a potential upcoming release or announcement.

09-30FLASHmodel_catalog

Two new model IDs, gemini-3.8-flash-lite-tts and gemini-3.8-flash-tts, have been added to the LiteLLM model catalog under Vertex AI.

09-30FLASHhuggingface

Google pushed new HuggingFace repositories for `tabfm-1.1.0` with JAX and PyTorch implementations.

09-30RTNgithub

Gemini CLI pre-release v0.64.0-nightly.20260930.g38700b4b3 includes fixes for autonomous plan execution, tool output truncation, and a2a-server settings migration.

09-29FLASHgithub

Gemini CLI released v0.63.0-preview.0 with fixes for connection recovery, memory optimization in agent loops, and autonomous plan execution, alongside a v0.62.0 release that includes a new "gemini 3.8 flash 3.5 flash lite" feature.

09-29RTNgithub

Google Gemini CLI published a new pre-release, v0.63.0-nightly.20260929.gfe6350238, which includes a fix for an infinite authentication loop.

09-28RTNgithub

Google Gemini CLI published a new pre-release, v0.63.0-nightly.20260928.g2fe7c2d3f.

10-02URGENTgithub_org

xAI released an official TypeScript SDK for their API, now publicly available on GitHub.

10-01URGENTgithub_org

xAI has made a new public repository called 'community-writer' available, which is described as a community-driven AI writer for Community Notes.

09-29FLASHmodel_catalog

Three new Xai Grok model IDs (global.xai.grok-4.7, us.xai.grok-4.7, xai.grok-4.7) appeared in the LiteLLM model catalog.

10-02RTNhacker news

DeepSeek's new "Harness" product is gaining significant traction on Hacker News.

09-29FLASHmodel_catalog

LiteLLM's model catalog now includes a new DeepSeek-V4.1-Flash-Fast model from Baseten, indicating potential support for a faster DeepSeek variant.

09-28FLASHmodel_catalog

Two new Deepseek models, prism/deepseek-v4-flash and prism/deepseek-v4.1-flash, have been added to the LiteLLM model catalog.

09-28FLASHlmarena

Two new DeepSeek models, 'deepseek-v4-pro-high-fireworks' and 'contenders/deepseek-v4-pro-high-fireworks', have appeared on the LMArena leaderboard.

09-28FLASHmodel_catalog

Two new Mistral OCR model IDs (2505 and 2512) detected in Azure AI's model catalog.

7 new model IDs detected across Azure AI, Cloudflare, Perplexity, and Strands, including 'kimi-k2-thinking' and 'pplx-decider-v1-27b'.

Aleph Alpha's new open-weight Kolibri LLM for German and English is gaining traction on Hacker News.

Three new model IDs (bespoke/bespokelabs/Bespoke-Nimble-9B, bespoke/nimble, bespoke/nimble-latest) have been added to the Litellm model catalog.

LiteLLM's model catalog now includes 'openrouter/inclusionai/ling-3.1-flash'.

OpenRouter API now lists 'nvidia/switchyard', suggesting a new unannounced model from Nvidia.

Microsoft pushed a new HuggingFace repository named 'microsoft/FrogNano-4B-2609'.

Three new 'laya' model IDs (laya/english, laya/multilingual, laya/typed-decisions) have appeared in the Litellm model catalog under 'other providers'.

An MIT Technology Review article titled "Don't be fooled–LLMs don't reason" is gaining significant traction on Hacker News, suggesting ongoing debate about AI capabilities.

Nvidia created a new HuggingFace repository called "nvidia/PixelDiT2-ImageNet". This is likely a new image generation model.

Nvidia created a new HuggingFace repository called 'nvidia/PixelUMM'.

Two new anonymous models, 'contenders/mimo-v2-6-flash-agent' and 'mimo-v2-6-flash-agent', have appeared on the LMArena leaderboard, potentially indicating unannounced model tests.

Two new anonymous models, "contenders/mimo-v2-6-pro-public" and "mimo-v2-6-pro-public," have appeared on the LMArena leaderboard, likely indicating unreleased stealth models.

Two new anonymous models, "barium-bb-v2-4xts-agent" and "contenders/barium-bb-v2-4xts-agent", appeared on the LMArena leaderboard, suggesting a potential stealth model release.

Two new anonymous models, "barium-bb-v2-4xts-webdev" and "barium-bb-wlqc-text," have appeared on the LMArena leaderboard, potentially indicating unannounced models undergoing testing.

A Hacker News story titled "Why Is Sam Altman a Free Man?" is gaining significant velocity, reaching 36 points per hour.

A new PDF paper alleges that AI companies are leaking user data to advertisers, gaining rapid traction on Hacker News.

NVIDIA created a new Hugging Face repository for a model called 'NVIDIA-Nemotron-Labs-3-Competitive-Coding-550B-A55B-NVFP4'.

Nvidia created two new HuggingFace repositories, 'Kumo-Relational' and 'Kumo-Tabular', which may indicate new models or datasets related to relational and tabular data processing.