AI Flash · live feed
A timeline of everything that has passed through the radar — newest first, with an evidence status for each item.
Continuously updated
Filter contains 225 items
How to read evidence statuses ↓
Starting with version 1.0.94-0, GitHub Copilot CLI discovers supported local models from a running Ollama service using the /model command. Once confirmed, they can be used without restarting. Selecting a local model does not enable offline mode or disable telemetry.
Falcon-ASR has 1.6 billion parameters and focuses on Arabic, particularly the Emirati dialect. According to Technology Innovation Institute, it achieved a WER of 20.92 % on six Arabic test sets. The available excerpt does not specify access conditions.
Hugging Face announced ongoing maintenance of the H200 infrastructure for the Inference Endpoints service, with completion expected around 12:30 UTC. The creation of new H200 instances may be temporarily limited.
Liquid AI has released the open weights of d1-3B and d1-omni-600M. The models return decisions in a single pass without generating tokens. They support text with images, and the experimental d1-omni-600M also supports text with audio.
Claude Haiku 5.5 is available through Claude API and four cloud offerings. It has a context window of one million tokens, a maximum output of 128 thousand tokens and support for adaptive thinking with the effort parameter.
According to the quoted announcement, the price of reading from the prompt cache for Claude Sonnet 5.5 has fallen from 0.20 to 0.10 USD per million tokens. The price of cache writes and all other prices remain unchanged.
The Python and TypeScript SDKs now offer beta classes for connecting the browser use and computer use tools to custom automation. The developer implements the tool methods, while the SDK handles the tool-calling loop.
The Claude Max and Claude Team plans now include monthly API credits. The announcement also links to instructions for obtaining them.
Claude Opus 3 remains available to all paying users of claude.ai after its retirement. Anthropic provides API access on request.
OpenAI updated the chat-latest snapshot, which points to the latest model available in ChatGPT on the Plus, Pro, Business and Enterprise plans. For production API use, it recommends the GPT-6 model family.
IDE updates are expected to restore accurate reporting of GitHub Copilot agent activity. The fix is available in Visual Studio Code from version 1.139.0; GitHub expects it for other IDEs by November 2026. Missing activity cannot be recovered retroactively.
Google released the open model EmbeddingGemma 2, which converts text, code, images, video and audio into a shared space of numerical vectors. It has 740 million parameters, uses the Apache 2.0 license and can run locally without an API key.
AWS has made it possible to create, configure, start and stop Amazon SageMaker Spaces environments directly through Amazon SageMaker Studio. It supports JupyterLab and Code Editor environments on Amazon SageMaker HyperPod EKS clusters; an administrator must prepare the add-on and permissions.
The Falcon-Emirati-7B model is based on the 7B variant of the Falcon-H1-Arabic family. According to its authors, it is intended to understand the Emirati dialect and generate text that takes the local cultural context into account. The available excerpt does not provide validation results or information on access options.
Gemini Nano Banana 2.1 is generally available. The update to Nano Banana 2 enables image generation and editing through conversation, supporting 1K, 2K and 4K resolutions as well as very wide and very tall formats.
OpenAI demonstrated a revamped Codex CLI tool in a video, featuring a full-screen interface and management of concurrent tasks. The demo includes launching three tasks by voice and transferring control to a phone. The available text does not specify the price, version or general availability.
The d1 model from Liquid AI processes both text and images and returns answer probabilities without generating tokens. It is available through the API and d1 Playground. The company lists a price of 0.04 USD per million input tokens and a text-based decision time of 200–300 ms.
According to OpenAI, administrators of eligible organizations can accept a standard Business Associate Agreement (BAA) and enable HIPAA compliance support directly in the API settings. Access depends on organizational eligibility.
The recommended replacement for Claude Opus 4.7 in GitHub Copilot is Claude Opus 5.5; the original announcement listed Claude Opus 5. The retirement of four models takes effect on 2 October 2026 across all environments of the service.
The AstaBrief 8B model generates scientific reports with citations and is available in the Fast mode of the Asta service. The authors are releasing both the weights and the training data. According to their measurements, the entire process takes an average of 51.1 seconds compared to 178.5 seconds in Thinking mode.
According to NVIDIA, a variant of the NVIDIA DGX Spark with 64 GB unified memory will be available this month from partner manufacturers. The text mentions the company Acer, but does not state the price or a specific sales start date.
The GPT-6 Astra Ultrafast model is available in OpenAI API and to eligible users of ChatGPT Work and Codex. It runs on NVIDIA Blackwell GPU and, according to NVIDIA, offers up to 8× faster token generation than Astra Standard mode.
The Live Data in Apps feature allows Amazon Quick applications to query Amazon Quick Sight datasets every time they are opened. According to AWS, the queries are executed with the user's permissions and respect row-level and column-level access restrictions.
The computer use feature is in public preview in the GitHub Copilot CLI and GitHub Copilot app tools for macOS and Windows. According to GitHub, it allows control of desktop applications. It requires user consent, and organizations can disable it.
AI Radar monitors Czech and international sources every day, looking for changes that truly deserve attention.
The result is a quick overview of what has actually changed in the AI world, rather than another stream of articles.
Use the CS/EN switch to read the same Radar in Czech or English. English content is published after its translation has been checked, so new and older items may appear later.
Each row represents one event — not one article. At a glance, you can see its significance, credibility and main point.
Illustrative example, not a current news item.
Agent mode is available to all paying users
Until now, the mode was available only on the highest plan; it is now available on all paid tiers without a waitlist.
The detail page contains a fuller summary, its significance and original sources. Practical impact appears in the detail and the For individuals and For businesses views. An AI Flash item reaches the main selection only after it has been expanded and meets the publication rules.
We first summarize each event in the same way for everyone. Based on those same facts, we then explain what the change means for your own use and what it could mean for how a company operates.
Today’s briefing is the same for everyone. Pages For individuals and For businesses can be found in the main navigation — they select only events relevant to the given use case.
Reset history · Outage history · Subscribe in a feed reader (RSS) · Methodology and label explanations
