# AI daily digest — 28 Sep 2026

Frontier models and agent safety led 11 AI releases on 28 September, with Anthropic introducing Claude Sonnet 5.5 and OpenAI publishing on safety cases for frontier training. Other launches include ElevenLabs' Eleven v4 speech model, xAI's Team Bots, H Company's Holo4 agents, NVIDIA's open agent safety platform and Meta's enterprise business AI platform.

Source-local date: Monday 28 September 2026 (America/Los_Angeles). 11 release(s). Canonical: https://mediarelease.co/digests/ai/2026-09-28/

## Releases

### [Sovereign AI: Full-Stack Infrastructure for AI Autonomy](https://mediarelease.co/us/mr01176-sovereign-ai-full-stack-infrastructure-for-ai-autonomy-28092026.html)

Issued by: SambaNova · Original source: https://sambanova.ai/blog/sovereign-ai

### [Towards safety cases for frontier AI training](https://mediarelease.co/us/mr01096-towards-safety-cases-for-frontier-ai-training-28092026.html)

Issued by: OpenAI · Original source: https://openai.com/index/towards-safety-cases-for-frontier-ai-training/

OpenAI published early guidelines for safety cases before continuing frontier reinforcement learning runs. The note covers technical safeguards such as alignment training, containment, and monitoring; operational practices; and investigations of misalignment incidents, with public disclosure expectations after reviews conclude.

### [See what 4 builders are making with Gemini 3.8 Flash](https://mediarelease.co/us/mr01080-see-what-4-builders-are-making-with-gemini-3-8-flash-28092026.html)

Issued by: Google · Original source: https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-flash-developers/

Google highlights four community projects built with Gemini 3.8 Flash, its latest workhorse model for software engineering and multistep reasoning. Examples include a live satellite-path simulation in Google Antigravity, an animated ink-style Seigaiha pattern, a prompted T. rex skeleton, and an interactive automatic-transmission model with multiple camera views.

### [Basis completes a tax workbook 2x faster with GPT-6 Astra](https://mediarelease.co/us/mr01079-basis-completes-a-tax-workbook-2x-faster-with-gpt-6-astra-28092026.html)

Issued by: OpenAI · Original source: https://openai.com/index/basis-tax-workbook-with-astra/

OpenAI reports that accounting automation firm Basis completed a 50-tab tax workbook about twice as fast with GPT-6 Astra as with GPT-5.6 Sol. Basis said the model better matched reasoning effort to each step, used tokens more efficiently, and lifted internal evaluation scores by about 20% through stronger intent understanding and fewer hand-written rules.

### [Team Bots: AI coworkers that learn from your team](https://mediarelease.co/us/mr01077-team-bots-ai-coworkers-that-learn-from-your-team-28092026.html)

Issued by: xAI · Original source: https://x.ai/news/team-bots

xAI launched Team Bots, shared Grok Bots that teams give files, apps, and expertise so everyone works from the same context. Each Team Bot combines context, plugins, credentials, and memories, with private per-user conversations and optional Slack handles. SpaceXAI uses them for account briefings, engineering coordination, brand review, and warehouse data questions; the feature is in public beta on Teams and Enterprise plans.

### [Introducing Claude Sonnet 5.5](https://mediarelease.co/us/mr01076-introducing-claude-sonnet-5-5-28092026.html)

Issued by: Anthropic · Original source: https://www.anthropic.com/claude-sonnet-5-5

Anthropic released Claude Sonnet 5.5, a faster and lower-cost model in its Claude 5.5 family for everyday coding and knowledge work. It scores much higher than Sonnet 5 on agentic coding tests, generates outputs over 30% faster, and costs $2/$10 per million input/output tokens. It is available on Amazon Web Services, Google Cloud, and Microsoft Azure with cyber and biology safeguards.

### [The Lenfest Institute grows landmark program with expanded OpenAI support](https://mediarelease.co/us/mr01075-lenfest-grows-landmark-program-with-openai-support-28092026.html)

Issued by: OpenAI · Original source: https://openai.com/index/lenfest-ai-collaborative-expansion/

OpenAI is expanding the Lenfest AI Collaborative and Fellowship Program with $5 million in funding plus up to $5 million in software credits and engineering support. The program embeds AI engineers in U.S. newsrooms; the next phase will add more organizations and turn fellowship projects into reusable tools for local journalism.

### [ElevenLabs Introduces Eleven v4 Text-to-Speech Model](https://mediarelease.co/us/mr01068-elevenlabs-introduces-eleven-v4-text-to-speech-model-28092026.html)

Issued by: ElevenLabs · Original source: https://elevenlabs.io/blog/eleven-v4

ElevenLabs launched Eleven v4, its most emotive text-to-speech model, plus a low-latency Eleven v4 Turbo variant for agents and real-time use. The company said v4 ranked first on Artificial Analysis and was preferred by about 75% of listeners in blind tests, with stronger tone, pacing, and voice cloning. Both models are live in ElevenAgents, ElevenCreative, and the API, with Turbo median inference latency around 100 milliseconds.

### [H Company Releases Holo4 Computer-Use Agent Models on Hugging Face](https://mediarelease.co/us/mr01067-h-company-releases-holo4-computer-use-agent-models-on-28092026.html)

Issued by: H Company · Original source: https://huggingface.co/blog/Hcompany/holo4

H Company released Holo4, a new series of agentic computer-use models available on Hugging Face and the H Models API in 27B dense and 35B-A3B mixture-of-experts sizes, plus Holotron4 Nano. Holo4 interacts with software through GUIs, code, MCP, and APIs and was trained with supervised and reinforcement learning for real business workflows. Model weights, trajectories, and a full collection in FP16, FP8, and GGUF formats were published alongside the announcement.

### [NVIDIA Launches Open Agent Safety Platform](https://mediarelease.co/us/mr01066-nvidia-launches-open-agent-safety-platform-28092026.html)

Issued by: NVIDIA · Original source: https://nvidianews.nvidia.com/news/open-agent-safety-platform

NVIDIA launched the Open Agent Safety Platform, an open software stack and reference design to secure AI agents from testing through deployment. It pairs OpenShell, an open-source secure runtime that traces agent actions on NVIDIA Vera CPUs, with Sentry, a BlueField-4 DPU watchdog that can quarantine out-of-bound agents in milliseconds. Partners named include Anthropic, Cisco, CrowdStrike, Microsoft, and others across infrastructure, models, and robotics.

### [Meta Launches Enterprise Platform for Business AI](https://mediarelease.co/us/mr01065-meta-launches-enterprise-platform-for-business-ai-28092026.html)

Issued by: Meta · Original source: https://about.fb.com/news/2026/09/launching-meta-enterprise-platform/

Meta launched Meta Enterprise Platform, a new business pillar to help companies use AI to grow, led by Chirantan "CJ" Desai as Chief Enterprise Platform Officer. The platform will bring Meta's models, agents, and infrastructure to businesses, starting with the Muse agent, Meta Business Agent, Muse API, and Muse Code. Desai joins from MongoDB, where he was CEO and President, and reports to Mark Zuckerberg.

More AI digests: https://mediarelease.co/digests/ai/
RSS: https://mediarelease.co/digests/ai/feed.xml