Frontier FluxFrontier AI
LatestOriginalsBriefArchiveVideosMethod
Frontier Flux hero: independent frontier AI coverage of models, benchmarks, labs, and chips

Independent · Capability-first

The latest AI news in realtime.

Evidence-led AI coverage across OpenAI, Anthropic, Google DeepMind, Meta, NVIDIA, xAI, and more.

Subscribe to daily brief

Free · Last 24 hours

Morning brief

Full brief →

Free independent brief of the latest, most significant AI news from the last 24 hours.

  1. 01
    Anthropic C.E.O. Dario Amodei Calls for A.I. Slowdown

    The New York Times reports: Anthropic C.E.O. Dario Amodei Calls for A.I. Slowdown.

  2. 02
    Grameenphone brings ChatGPT, Gemini, Claude and Grok to MyGP

    The Business Standard reports: Grameenphone brings ChatGPT, Gemini, Claude and Grok to MyGP.

Get it by email

Free daily email: model releases, benchmarks, and lab notes from the last 24 hours. OpenAI, Anthropic, DeepMind, Meta, NVIDIA, and more.

Opens Buttondown to finish signup. Confirm by email if asked. No ads. Unsubscribe anytime.

LIVEUpdated 6h ago
AllOriginalOpenAIAnthropicGoogleMetaNVIDIAMSFTxAIMistralOther
LeadAnthropicGoogleOpenAI6h ago
Anthropic C.E.O. Dario Amodei Calls for A.I. Slowdown

Anthropic C.E.O. Dario Amodei Calls for A.I. Slowdown

The New York Times reports: Anthropic C.E.O. Dario Amodei Calls for A.I. Slowdown.

viaThe New York Times+Moneycontrol.com+Hindustan Times+lokmattimes.com+The Tribune+LinkedIn+Gizmodo+dw.com+Forbes+BBC

Originals

All originals

Frontier Flux field notes and checklists: capability and practice, not company soap opera.

  • Abstract dark field with cyan loop paths and nested session rings, AI-generated
    Original35h ago· 2 min read

    OpenAI puts the Codex harness on Agents API public beta

    On 10 September 2026 OpenAI Developers said the Agents API is in public beta: a managed Codex harness for cloud agents, with OpenAI handling orchestration, long-running sessions, and context management. OpenAI says there is no extra Agents API fee beyond tokens and tools.

  • Abstract cyan rings and stacked planes on a dark field, AI-generated
    Original2d ago· 2 min read

    DeepSeek ships V4.1-Flash MIT weights and API deepseek-flash

    DeepSeek released DeepSeek-V4.1-Flash with MIT weights on Hugging Face and API id deepseek-flash. The lab describes a 552B MoE Causal Encoder-Decoder that activates 8B parameters on prefill and 16B on decode, plus a 890-byte global KV cache per token.

  • Abstract dark field with a cyan flare path and amber sunburst lattice, AI-generated
    Original3d ago· 2 min read

    OpenAI ships ChatGPT Images 2.5 with API Flare and Sunburst

    On 8 September 2026 OpenAI said ChatGPT Images 2.5 is rolling out to all ChatGPT, ChatGPT Work, and Codex users, with two API models: GPT-Image-2.5 Flare and GPT-Image-2.5 Sunburst. The lab reports up to 50% lower generation latency versus Images 2.0, plus Sketch, templates, and on-image comments in ChatGPT.

Selected videos

All selected videos

Official lab and research clips tied to stories when we can. Selected coverage, not a channel firehose.

  • AI researchers debate how close we are to recursive self-improvement

    Dwarkesh long-form interview relevant to AI.

    Dwarkesh Patel · 37h ago · Open on YouTube

  • AI News: The AI World is REALLY Scared Right Now

    Matt Wolfe product/tooling pass on AI.

    Matt Wolfe · 39h ago · Open on YouTube

Wire· 20 stories (lead + 19)

One slot per story cluster. Multi-source items show an N sources badge. Capability and artifact first, not a raw ticker.

Sort by
Most recentOldest first
OpenAIAnthropicGooglexAI14h ago· The Business Standard

Grameenphone brings ChatGPT, Gemini, Claude and Grok to MyGP

The Business Standard reports: Grameenphone brings ChatGPT, Gemini, Claude and Grok to MyGP.

Preview: Grameenphone brings ChatGPT, Gemini, Claude and Grok to MyGP
OriginalOpenAI35h ago· 2 min read

OpenAI puts the Codex harness on Agents API public beta

On 10 September 2026 OpenAI Developers said the Agents API is in public beta: a managed Codex harness for cloud agents, with OpenAI handling orchestration, long-running sessions, and context management. OpenAI says there is no extra Agents API fee beyond tokens and tools.

Abstract dark field with cyan loop paths and nested session rings, AI-generated
OriginalOther2d ago· 2 min read

DeepSeek ships V4.1-Flash MIT weights and API deepseek-flash

DeepSeek released DeepSeek-V4.1-Flash with MIT weights on Hugging Face and API id deepseek-flash. The lab describes a 552B MoE Causal Encoder-Decoder that activates 8B parameters on prefill and 16B on decode, plus a 890-byte global KV cache per token.

Abstract cyan rings and stacked planes on a dark field, AI-generated
OriginalOpenAI3d ago· 2 min read

OpenAI ships ChatGPT Images 2.5 with API Flare and Sunburst

On 8 September 2026 OpenAI said ChatGPT Images 2.5 is rolling out to all ChatGPT, ChatGPT Work, and Codex users, with two API models: GPT-Image-2.5 Flare and GPT-Image-2.5 Sunburst. The lab reports up to 50% lower generation latency versus Images 2.0, plus Sketch, templates, and on-image comments in ChatGPT.

Abstract dark field with a cyan flare path and amber sunburst lattice, AI-generated
OriginalOpenAI4d ago· 2 min read

OpenAI reports a Navier-Stokes singularity proof from an unnamed next-gen model

On 8 September 2026 OpenAI published a research note saying a group of agents, using an unnamed internal model it describes as significantly more capable than GPT-6 Astra, produced an analytical proof and a Lean formalization that three-dimensional incompressible Navier-Stokes can develop a singularity in finite time. The company says it does not intend to claim the Clay Millennium Prize.

Abstract teal and amber vortex spiraling over a dark navy field
OriginalOpenAI5d ago· 2 min read

Pachocki: no lab is ready to scale at maximum speed

On 6 September 2026 OpenAI chief scientist Jakub Pachocki published An Alien Mind. He writes that no lab has solved alignment and monitoring enough to keep scaling at maximum speed for much longer, and he hopes voluntary slowdowns become common until shared safety bars exist.

Abstract dark field with a cyan lattice and an amber restraint bar, AI-generated
OriginalOpenAI8d ago· 3 min read

OpenAI rolls GPT-6 Astra to Pro, Enterprise, and the API

On 4 September 2026 OpenAI said GPT-6 Astra is now available to Pro, Enterprise, and Business Premium users in ChatGPT Work and Codex, and live in the API. Plus and remaining Business users might wait a few more days. The lab calls Astra its most capable broadly deployed model and its first Critical cybersecurity model under the Preparedness Framework.

Abstract dark field with cyan access lanes and a gated amber lattice, AI-generated
OriginalOther9d ago· 2 min read

IFM ships K2 Horizon: six open-weight models, 0.9B to 375B-A23B

On 3 September 2026 the Institute of Foundation Models released K2 Horizon, a six-model fleet from 0.9B to 375B-A23B, with Apache 2.0 Hub tags. IFM says the 0.9B, 3.7B, and 7B models set size-class state of the art. The 375B card says the final checkpoint is live while some data and training code still will be released.

Abstract dark field with cyan nested scale planes and faint routing arcs, AI-generated
OriginalGoogle10d ago· 3 min read

Google ships Gemini 3.8 Flash at 3.7 prices; Flash Cyber stays Fairwind-only

On 2 September 2026 Google launched Gemini 3.8 Flash as its new workhorse at the same introductory $0.75/$3.75 per million tokens as 3.7 Flash, plus Gemini 3.8 Flash Cyber for trusted defenders through the new Fairwind Program.

Abstract dark field with cyan agentic loops and an amber gated lattice, AI-generated
OriginalMeta11d ago· 2 min read

Meta launches Muse Voice Transcribe: streaming ASR and 20-plus-speaker diarization

On 1 September 2026 Meta Superintelligence Labs launched Muse Voice Transcribe, its first real-time audio perception model, with streaming ASR, 20-plus-speaker diarization, and endpointing. It is live on the Meta Model API, Meta AI for Mac, and Muse Code. Rankings Meta cites are as of that date.

Abstract dark field with cyan audio waveforms splitting into stacked speaker ribbons, AI-generated

Showing 10 of 19 stories

Frontier Flux Originals from this window also live on Originals. Stories older than 30 days move to the Archive. Search still finds them.

From X

Lab and researcher posts as receipts for the wire, not the primary headline stream.

“It's now clear that alignment is critical and won't be solved behind the closed doors of a handful of frontier labs. So today we're launching the Open Alignment Initiative, led by @Thom_Wolf @huggingface and asking to be part of the "embedded evaluators" program that...”

15h ago·Clem Delangue on X·Loading embed…

“GPT-Live in the API, for empowering builders to create new kinds of applications”

2d ago·Greg Brockman on X·Loading embed…

“we used our models to find and fix critical vulnerabilities in our own systems, as part of a 250+ person effort within openai focused on using AI for cyber defense and offense research”

3d ago·Greg Brockman on X·Loading embed…

“the early usage on muse has blown way past our projections. muse's users today are using 10x more than our testing cohorts. enjoy!”

4d ago·@alexandr_wang on X·Loading embed…

“we’re proud to be partnering with @1Password so Muse will be able to work with the logins you already have. more details on how we built all of this here:”

4d ago·@alexandr_wang on X·Loading embed…

Frontier Flux is independent, capability-first coverage of frontier AI: model releases, benchmarks, papers, chips, and attributed lab statements. Not affiliated with key AI figures, OpenAI, Anthropic, Google, Meta, NVIDIA, Microsoft, xAI, Mistral, or X. Wire cards link out to original reporting. Frontier Flux originals host a full story on-site and still cite the primary post.

OriginalsBriefArchiveVideosSearchMethodRSSLatest

Thumbnails and previews are used under fair use for news aggregation and link to original sources.