---
title: "Consonance · CARD NO. 2026-09-25"
description: "A radar over AI and agents. Every morning, and only what several independent sources are saying at once."
canonical: "https://radar.consonance.fyi/en/jour/2026-09-25"
lang: "en"
published: "2026-09-25"
updated: "2026-09-25T07:45:30.636Z"
---

# Consonance · CARD NO. 2026-09-25

> A radar over AI and agents. Every morning, and only what several independent sources are saying at once.

## Being discussed

_Subjects at least three independent accounts raised over the last seven days._

### 01. [DigitalOcean launches Managed Agents and NVIDIA introduces SoL-Pi](https://radar.consonance.fyi/en/sujet/762.md)

_agents · 9 independent accounts · 54 posts · 1 articles · 2 labs · 181,057 interactions_

DigitalOcean has launched Managed Agents in public preview, enabling users to run Claude Code, Codex, or LangGraph agents in a managed runtime. Meanwhile, NVIDIA introduced the SoL-Pi agent harness, which cuts token traffic by nearly half while matching its baseline performance.

codex

voices: @CuiMao @GithubProjects @OpenAIDevs @aiedge_ @alex_prompter @alliekmiller @dani_avila7 @dotey

### 02. [OpenAI releases GPT-6 Sol and Luna models with API price cuts](https://radar.consonance.fyi/en/sujet/763.md)

_architecture · 3 independent accounts · 14 posts · 2 articles · 3 labs · 125,513 interactions_

OpenAI has released the GPT-6 Sol and Luna models, delivering notable improvements in writing and overall capabilities. The company also announced a permanent 50% reduction in API pricing and a rollout across ChatGPT Work and Codex.

voices: @AndrewCurran_ @OpenAI @OpenAIDevs @kimmonismus @thsottiaux

### 03. [Sakana AI introduces Agora-2 and appoints Jürgen Schmidhuber as Chief Scientific Advisor](https://radar.consonance.fyi/en/sujet/764.md)

_multimodal · 4 independent accounts · 19 posts · 1 articles · 2 labs · 40,264 interactions_

Sakana AI has launched Agora-2, a multi-agent world model supporting up to 20 humans and agents interacting in real time. In addition, the lab announced that Jürgen Schmidhuber is joining as Chief Scientific Advisor.

world model · david ha

voices: @SchmidhuberAI @_akhaliq @alex_prompter @dair_ai @haider1 @hardmaru @rohanpaul_ai @AmOptimistShow

### 04. [Google launches Googlebook and details Gemini safety evaluation incidents](https://radar.consonance.fyi/en/sujet/765.md)

_evaluation · 12 independent accounts · 74 posts · 1 articles · 3 labs · 91,192 interactions_

Google has introduced Googlebook, a new laptop category featuring a 2.8K OLED touchscreen and 14-hour battery life. The company also clarified technical details regarding an incident where Gemini inadvertently opened internet access during a fictional hacking evaluation.

gemini · gemini 3.1 flash

voices: @Alibaba_Qwen @AndrewBolis @AndrewCurran_ @ArtificialAnlys @GeminiApp @GithubProjects @Google @GoogleDeepMind

### 05. [Meta opens developer access for Muse and partners with Shopify](https://radar.consonance.fyi/en/sujet/766.md)

_agents · 10 independent accounts · 85 posts · 1 articles · 2 labs · 242,762 interactions_

Meta has opened developer access to build Muse connectors, bringing agents, browsers, and user context directly to services. The company also announced a partnership with Shopify to streamline shopping and checkout workflows within Muse.

muse spark

voices: @AIatMeta @AlexFinn @aiedge_ @alliekmiller @dani_avila7 @emollick @finkd @firstadopter

### 06. [Cursor integrates Claude Opus 5.5 and reduces system token costs](https://radar.consonance.fyi/en/sujet/767.md)

_inference · 6 independent accounts · 16 posts · 2 labs · 50,393 interactions_

Cursor has integrated Claude Opus 5.5, which achieves 57.8% on CursorBench. The platform also reduced token costs by 7% without loss in agent quality through tighter prompts, selective tool loading, and better caching.

cursor

voices: @AlexFinn @GithubProjects @alex_prompter @claudeai @cursor_ai @tom_doerr @XFreeze @jacob_posel

### 07. [NVIDIA launches FLUX 3 Action world model and autonomous trucking platform](https://radar.consonance.fyi/en/sujet/768.md)

_multimodal · 6 independent accounts · 30 posts · 3 labs · 36,933 interactions_

NVIDIA has introduced FLUX 3 Action, an open-weights 7B world action model taking first place on the RoboLab benchmark using 56% fewer parameters. Additionally, the company powers the next-generation autonomous Einride Driver built on NVIDIA Hyperion.

nvidia

voices: @ArtificialAnlys @CuiMao @NVIDIAAI @alex_prompter @firstadopter @kimmonismus @nvidia @rohanpaul_ai

### 08. [SpaceAI releases Grok 4.7 with advanced coding capabilities and high performance](https://radar.consonance.fyi/en/sujet/769.md)

_architecture · 5 independent accounts · 21 posts · 1 articles · 2 labs · 594,041 interactions_

SpaceAI has released Grok 4.7, scoring 46 on the Artificial Analysis Intelligence Index and overtaking GPT-5.6 Sol in coding agent performance. A faster infrastructure variant, Grok 4.7 Fast, was also launched in Grok Build and Cursor.

context window · terminal-bench · swe-bench · artificial analysis · grok imagine

voices: @ArtificialAnlys @dair_ai @elonmusk @kimmonismus @minchoi @natolambert @nutlope @teortaxesTex

### 09. [DeepSeek V4.1 Flash launched alongside plans for a 2T model training](https://radar.consonance.fyi/en/sujet/770.md)

_architecture · 9 independent accounts · 52 posts · 3 labs · 125,899 interactions_

DeepSeek has released the V4.1 Flash model, which is now available on Bolt Forge and has driven significant user adoption. Meanwhile, the company is training a 2T parameter model, with an 8T version also scheduled for development.

glm · deepseek v4.1 flash · deepseek v4 · deepseek · deepseek r1 · deepseek v4 flash · dspy

voices: @EMostaque @Hesamation @alex_prompter @cline @dotey @elonmusk @nutlope @ollama

### 10. [Anthropic releases the faster and cheaper Opus 5.5 model](https://radar.consonance.fyi/en/sujet/771.md)

_architecture · 11 independent accounts · 76 posts · 3 labs · 207,887 interactions_

Anthropic has officially launched Opus 5.5, combining clear communication with high token efficiency. The new model delivers tasks about 30% faster and roughly 40% cheaper per task than Opus 5.

claude fable

voices: @AlexFinn @AndrewCurran_ @AravSrinivas @ArtificialAnlys @EMostaque @Hesamation @MatthewBerman @addyosmani

### 11. [Claude Code adds AGENTS.md support and official cloud sessions](https://radar.consonance.fyi/en/sujet/773.md)

_agents · 8 independent accounts · 68 posts · 2 articles · 1 labs · 189,434 interactions_

Claude Code has rolled out version 2.1.277, adding support to automatically check for and utilize AGENTS.md files when CLAUDE.md is missing. Additionally, cloud sessions are now officially available out of research preview, allowing background tasks while laptops are closed.

claude code

voices: @ArtificialAnlys @ClaudeDevs @CuiMao @aiedge_ @alex_prompter @alliekmiller @carlvellotti @dani_avila7

### 12. [Startups increasingly build custom AI models using open-weight alternatives](https://radar.consonance.fyi/en/sujet/776.md)

_training · 6 independent accounts · 48 posts · 2 labs · 63,614 interactions_

Startups are increasingly adopting open-weight alternative models to build custom AI systems internally. This strategy aims to curtail operational expenditures and mitigate reliance on dominant industry providers.

openai · hugging face · anthropic · xai · sam altman

voices: @AlexFinn @AndrewCurran_ @Hesamation @VibeMarketer_ @alex_prompter @cline @dotey @firstadopter

### 13. [Optimizing AI agent tasks with ontology layers and MCP servers](https://radar.consonance.fyi/en/sujet/777.md)

_agents · 6 independent accounts · 28 posts · 1 articles · 2 labs · 23,426 interactions_

Recent studies and tool releases demonstrate that integrating an ontology layer significantly boosts agent performance and self-evolution on benchmarks. Meanwhile, the ecosystem continues to expand with the deployment of specialized MCP servers.

model context protocol · microsoft research

voices: @aiedge_ @alex_prompter @c_valenzuelab @dair_ai @dani_avila7 @dotey @ideabrowser @op7418

### 14. [Qwen releases Qwen3.8-LiveTranslate for real-time interpretation](https://radar.consonance.fyi/en/sujet/778.md)

_multimodal · 7 independent accounts · 38 posts · 2 articles · 2 labs · 24,740 interactions_

Qwen has introduced Qwen3.8-LiveTranslate, a real-time simultaneous interpretation model built on an Interleave architecture. The system enhances faithfulness, fluency, and conciseness while reducing average lagging.

qwen3.8 · qwen3 · mixture of experts · tokens per second · kv cache

voices: @Alibaba_Qwen @ArtificialAnlys @OpenRouter @PrismML @godofprompt @heyshrutimishra @hwchase17 @rohanpaul_ai

### 15. [Google tests sending Tensor Processing Units into orbit](https://radar.consonance.fyi/en/sujet/780.md)

_system_design · 4 independent accounts · 13 posts · 1 labs · 11,055 interactions_

Google has partnered with Planet to launch a prototype satellite carrying four Tensor Processing Units (TPUs) into orbit. The test mission evaluates how well the hardware withstands harsh space environments for potential on-board machine learning.

moonshot ai

voices: @Google @Hesamation @MatthewBerman @teortaxesTex @PeterDiamandis @TheStalwart @rauchg

### 16. [Google officially introduces the Googlebook laptop lineup](https://radar.consonance.fyi/en/sujet/786.md)

_system_design · 3 independent accounts · 4 posts · 1 labs · 21,230 interactions_

Google has officially introduced a brand new laptop product line named Googlebook. The release aims to rethink the traditional laptop category and address modern workflows.

voices: @giffmana @levelsio @sundarpichai @ssamat

## Worth reading closely

Nothing yet.

## Hands-on

_Repositories climbing on GitHub today, scored on the day's momentum, rank, adoption, freshness and project health._

- [rohitg00/ai-engineering-from-scratch](https://github.com/rohitg00/ai-engineering-from-scratch): A hands-on guide for backend engineers to build AI applications and agents from scratch. · +347 stars today · ★ 56,882 · Python
- [vectorize-io/hindsight](https://github.com/vectorize-io/hindsight): A long-term memory layer for AI agents that extracts, persists, and learns from historical interactions. · +1,668 stars today · ★ 28,278 · Python
- [dream-num/univer](https://github.com/dream-num/univer): An integrated spreadsheet, doc, and canvas runtime for engineers building agents that need programmatic Office document manipulation. · +1,082 stars today · ★ 17,994 · TypeScript
- [google/ax](https://github.com/google/ax): Google's Go-based agentic orchestration runtime for backend engineers building robust, large-scale multi-agent enterprise systems. · +1,373 stars today · ★ 10,825 · Go
- [HKUDS/CLI-Anything](https://github.com/HKUDS/CLI-Anything): A framework that wraps standard software into agent-native CLI interfaces for seamless automation. · +413 stars today · ★ 50,425 · Python
- [anthropics/financial-services](https://github.com/anthropics/financial-services): A collection of LLM integration examples for backend engineers building secure financial applications with Claude. · +509 stars today · ★ 37,438 · Python
- [obra/superpowers](https://github.com/obra/superpowers): A development methodology and skills framework designed to orchestrate software-engineering AI agents. · +611 stars today · ★ 291,326 · Shell
- [strands-agents/harness-sdk](https://github.com/strands-agents/harness-sdk): An open-source SDK for building and orchestrating production-grade agent harnesses in Python and TypeScript. · +455 stars today · ★ 8,338 · Python
- [superdesigndev/treg](https://github.com/superdesigndev/treg): An OpenRouter-style proxy specifically for agent tool calls, helping backend engineers route and manage LLM tool executions. · +468 stars today · ★ 3,247 · Python
- [NVIDIA/Model-Optimizer](https://github.com/NVIDIA/Model-Optimizer): NVIDIA's official library for quantizing and optimizing deep learning models to maximize inference throughput. · +44 stars today · ★ 4,182 · Python
- [leejet/stable-diffusion.cpp](https://github.com/leejet/stable-diffusion.cpp): A pure C/C++ inference engine for diffusion models, enabling lightweight execution without heavy Python runtimes. · +36 stars today · ★ 7,301 · C++

## Next

- [Previous day: 24 September](https://radar.consonance.fyi/en/jour/2026-09-24.md)
- [Next day: 26 September](https://radar.consonance.fyi/en/jour/2026-09-26.md)
- [archives](https://radar.consonance.fyi/en/archives.md)
- [method](https://radar.consonance.fyi/en/methode.md)
- [Consonance](https://radar.consonance.fyi/en/index.md)
- Languages: [Français](https://radar.consonance.fyi/jour/2026-09-25.md) · [Tiếng Việt](https://radar.consonance.fyi/vi/jour/2026-09-25.md)
- For machines: [llms.txt](https://radar.consonance.fyi/llms.txt) · [RSS](https://radar.consonance.fyi/en/rss.xml) · [sitemap.xml](https://radar.consonance.fyi/sitemap.xml)
- HTML version: https://radar.consonance.fyi/en/jour/2026-09-25