Nemotron 3 Super Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning paper: https://
huggingface.co/papers/2604.12
374
…
LLMS
-

Nemotron 3 Super: Open MoE Hybrid Model for Agentic AI
By
–
-

Google Gemini 3.1 Flash Shows Strong Progress on TTS
By
–
The progress from 2.5 to 3.1 has been super strong! Excited to see this much improvement for a Flash model. Learn more in our blog: https://
blog.google/innovation-and
-ai/models-and-research/gemini-models/gemini-3-1-flash-tts/
… -

Gemini 3.1 Flash TTS: Advanced Text-to-Speech Model Launches
By
–
Introducing Gemini 3.1 Flash TTS 🗣️, our latest text to speech model with scene direction, speaker level specificity, audio tags, more natural + expressive voices, and support for 70 different languages.
— Logan Kilpatrick (@OfficialLoganK) 15 avril 2026
Available via our new audio playground in AI Studio and in the Gemini API! pic.twitter.com/5PpBdhQMNgIntroducing Gemini 3.1 Flash TTS , our latest text to speech model with scene direction, speaker level specificity, audio tags, more natural + expressive voices, and support for 70 different languages. Available via our new audio playground in AI Studio and in the Gemini API!
-
Databricks AI Gateway Extends Unity Catalog Governance to Agentic AI
By
–
AI Gateway in Databricks now extends Unity Catalog governance to agentic AI.
— Databricks (@databricks) 15 avril 2026
As agents call LLMs, pull data through MCP servers, and invoke external APIs, every step touches sensitive data and audit requirements. With the latest updates to AI Gateway, you can now bring all of… pic.twitter.com/WICCgaXxiEAI Gateway in Databricks now extends Unity Catalog governance to agentic AI. As agents call LLMs, pull data through MCP servers, and invoke external APIs, every step touches sensitive data and audit requirements. With the latest updates to AI Gateway, you can now bring all of
-
Internal AI Models Far Ahead of Public Releases
By
–
We don't have Mythos. The internal models of Google, OpenAI and Anthropic are substantially ahead of what the public gets to use. The important question is whether dissemination stops.
-
Zero-Latency Inference Transforms Engineer Workflows at OpenAI
By
–
"I used a billion tokens this week. I'm not even in the top 100 Codex users at OpenAI."
— Cerebras (@cerebras) 15 avril 2026
We sat down with @jxnlco (creator of Instructor, now on @OpenAI's Developer Experience team) to talk about how zero-latency inference is changing the way engineers work. pic.twitter.com/PDyHtYRXec"I used a billion tokens this week. I'm not even in the top 100 Codex users at OpenAI." We sat down with @jxnlco (creator of Instructor, now on @OpenAI
's Developer Experience team) to talk about how zero-latency inference is changing the way engineers work. -

Rethinking On-Policy Distillation of Large Language Models Research
By
–
Rethinking On-Policy Distillation of Large Language Models Phenomenology, Mechanism, and Recipe paper: https://
huggingface.co/papers/2604.13
016
… -

KnowRL Boosts LLM Reasoning via Reinforcement Learning with Knowledge Guidance
By
–
KnowRL Boosting LLM Reasoning via Reinforcement Learning with Minimal-Sufficient Knowledge Guidance paper: https://
huggingface.co/papers/2604.12
627
… -
ClawGUI: Unified Framework for Training and Deploying GUI Agents
By
–
ClawGUI
— AK (@_akhaliq) 15 avril 2026
A Unified Framework for Training, Evaluating, and Deploying GUI Agents
paper: https://t.co/VPMTifMWGK pic.twitter.com/xISqw6h3pQClawGUI A Unified Framework for Training, Evaluating, and Deploying GUI Agents paper: https://
huggingface.co/papers/2604.11
784
… -
Training LLMs on AI outputs risks model degradation
By
–
What can happen when LLM models are trained on AI outputs? A potential hidden, bad influence https://
nature.com/articles/d4158
6-026-00906-0
…