AI news today

“PhysBrain 1.5 unifies physical environment understanding, action generation, and future state prediction via joint autoregressi…”

They claim — Across 28 embodied understanding benchmarks, our 8B model achieves an average score of 72.5, setting a new open-source state of the art and performing on par with leading proprietary mode…

Published 2026-09-13

Fetched 2026-09-15 23:01 UTC · 70061e1a1aeb68ea

Torn from — the abstract

“Dream-RSI enables scalable recursive self-improvement by using historical discovery replay to evaluate exploration policies off…”

Published 2026-09-13

Fetched 2026-09-15 23:01 UTC · 263aacf523ded915

Torn from — the abstract

“Atria Dawn Preview is a foundation agentic language model trained through verified tool interactions that achieves strong bench…”

They claim — Across 16 benchmarks spanning real-world research, engineering, and digital work, Atria Dawn Preview is competitive with frontier agents and achieves the highest reported score on five of…

Published 2026-09-13

Fetched 2026-09-15 23:01 UTC · 21c391716120f827

Torn from — the abstract

This source gives a title and a link, nothing to quote.

Published 2026-09-13

Fetched 2026-09-13 23:03 UTC · df33ebe0d318324d

Torn from — the page itself

“Official Project Page for Recurrent Looped Transformer (RLT)”

They claim — Training length is 32 operations; evaluation extends to 128 operations (4× the training length), with 2,048 test programs per task and length.

Published 2026-09-12

Fetched 2026-09-15 23:01 UTC · 675776a9a37fe26a

What the account wrote — The account left it blank

Torn from — the README

“Realtime spoken interaction demands deep reasoning, prompt responses, and fluid turn-taking. We present StepAudio 3 Realtime, a…”

They claim — StepAudio 3 Realtime achieves top-tier performance across key dimensions: an exceptional 90.6 on the MMSU benchmark, 98.9 Overall on the Artificial Analysis Full-Duplex Bench, and a 56.0%…

Published 2026-09-11

Fetched 2026-09-16 23:01 UTC · c13c53fb82422478

Torn from — the abstract

“Free open-source extractor for AI coding assistant chat histories. Supports Claude Code, Cursor, Windsurf, Aider, Cline/Roo Cod…”

Published 2026-09-11

Fetched 2026-09-15 23:01 UTC · 32cd43375b85b5e6

What the account wrote — The account left it blank

Torn from — the README

This source gives a title and a link, nothing to quote.

They claim — (<1% loss) and as a result getting a x1.95 speed-up on several tasks.

Published 2026-09-11

Fetched 2026-09-15 23:01 UTC · fae0d09608b75058

Torn from — the model card

This source gives a title and a link, nothing to quote.

Published 2026-09-11

Fetched 2026-09-15 23:01 UTC · 65418212282ccac7

Torn from — the model card

This source gives a title and a link, nothing to quote.

They claim — The refusal circuitry is surgically removed while every capability-critical component (routed experts, Engram memory, CSA2 sparse attention, DSpark draft head, vision tower, router gates,…

Published 2026-09-10

Fetched 2026-09-13 23:03 UTC · aba564a993aae2f3

Torn from — the model card

“LIT improves robot action generalization by first training pose-conditioned action priors without images, then constraining vis…”

They claim — Across four vision-language-action and world-action architectures (Pi0.5, MolmoAct2, FAST-WAM, and ImageWAM), LIT improves overall LIBERO-Plus success by 3.87-10.70 percentage points whil…

Published 2026-09-10

Fetched 2026-09-14 23:01 UTC · 3e83d75847ca11ec

Torn from — the abstract

“ZGCM-1 is a 7B open foundation model that combines internal reasoning with external tool use, trained via efficient architectur…”

They claim — On several challenging mathematical reasoning and agentic search suites, it remains competitive with frontier models orders of magnitude larger, such as Qwen3-235B-A22B and GLM-5.1.

Published 2026-09-10

Fetched 2026-09-15 23:01 UTC · fcb1e94d9a32e870

Torn from — the abstract

This source gives a title and a link, nothing to quote.

They claim — TWIN TURBO: Closed source level of intelligence with an arc-c at 709 (this is instruct "medium", reasoning is even higher), while using smaller quants too.

Published 2026-09-10

Fetched 2026-09-14 23:01 UTC · 68058414275bc9a3

Torn from — the model card

“Benchmark Radar is a searchable living database and discovery engine for AI evaluation benchmarks that aggregates sources, scor…”

Published 2026-09-09

Fetched 2026-09-14 23:01 UTC · 1cc8c83680c0c992

Torn from — the abstract

“Vidu S2 introduces real-time interactive avatar and video editing models that support high-resolution spatial video generation …”

They claim — Experiments show that Vidu S2 outperforms all baselines.

Published 2026-09-09

Fetched 2026-09-15 23:01 UTC · 5e45abe94152d970

Torn from — the abstract

This source gives a title and a link, nothing to quote.

They claim — (<1% loss) and as a result getting a x1.95 speed-up on several tasks.

Published 2026-09-08

Fetched 2026-09-14 23:01 UTC · b4d8039f21680101

Torn from — the model card

“A data-centric framework with specialized systems for reasoning analysis, cost reduction, and execution verification enables sm…”

They claim — The three checkpoints improve over their starting models by an average of 23.76% on the full CyberGym suite and 10.49% across the pooled CTF suites.

Published 2026-09-07

Fetched 2026-09-14 23:01 UTC · 4d29dafb2e864139

Torn from — the abstract

“A CLI that reads Claude Code, Codex, and Gemini CLI session logs and works out how much they cost, by model, project, and day.”

They claim — Cache hit under ~80%, consistently - your context probably isn't stable between calls, or caching isn't switched on.

Published 2026-09-07

Fetched 2026-09-13 23:03 UTC · 285e008f28d1dc05

What the account wrote — Newcastle

Torn from — the README

“Composing complementary continual learning mechanisms substantially improves long-horizon memorization in language models by co…”

They claim — Our best method combines all three anchors with merged LoRA, ranks among the top 3 methods in all datasets, and raises average final retention from 1.2% under naive sequential fine-tuning…

Published 2026-09-06

Fetched 2026-09-16 23:01 UTC · 299fad04b14317fa

Torn from — the abstract

“DLSS 5 NR on your whole Windows desktop, in real time with Boost, FG, etc. RTX 30/40/50, 12 languages, user presets, recording …”

They claim — Windows 10 and multi-GPU systems are experimental — v1.12 fixes adapter/output selection from user logs, not local hardware.

Published 2026-09-06

Fetched 2026-09-15 23:01 UTC · b6fda159d2186867

What the account wrote — The account left it blank

Torn from — the README

“Professional agent skills for screenwriting, television writing and dramaturgy”

Published 2026-09-06

Fetched 2026-09-15 23:01 UTC · 23ae32d1655cbeed

What the account wrote — Toronto, Canada

Torn from — the README

“DataFlex-RL evaluates reinforcement learning data policies and finds that uniform sampling matches or exceeds adaptive rollout …”

They claim — Uniform GRPO improves the domain-balanced average accuracy by 7.76 percentage points over the untrained checkpoint.

Published 2026-09-04

Fetched 2026-09-14 23:01 UTC · 8b008970febe99b9

Torn from — the abstract

This source gives a title and a link, nothing to quote.

They claim — Within the 10B-scale general-purpose VLMs compared in this release blog, ZDTaichu5.0-9B retains first-tier general visual understanding while supporting spatial reasoning, high-level embo…

Published 2026-09-04

Fetched 2026-09-15 23:01 UTC · 5179574dcefd9e42

Torn from — the model card

“An agent-first linter for Tailwind design systems. Write design system rules that agents can verify.”

They claim — 10% to 48% less than with rules alone.

Published 2026-09-02

Fetched 2026-09-15 23:01 UTC · 2bae6979048769f9

What the account wrote — The account left it blank

Torn from — the README

“One-click setup of the leaked DLSS 5 neural-rendering build for any DX11/DX12 game on RTX 20–50, with or without DLSS. ReShad…”

They claim — exe) - the game exe is detected automatically (the folder and two levels below it are searched, so bin\x64dx12\ and Unreal Binaries\Win64\ layouts work; Unity crash handlers, Unreal helpe…

Published 2026-08-31

Fetched 2026-09-13 23:03 UTC · 38bc58445c3c8fae

What the account wrote — The account left it blank

Torn from — the README

This source gives a title and a link, nothing to quote.

Published 2024-07-18

Fetched 2026-09-14 23:01 UTC · bad7065e70b23d0e

Torn from — the listing only (the original would not open)

“Introducing ChatGPT Images 2.5 OpenAI's image generation models are apparently used "more than 3 billion images across ChatGPT …”

Published 2026-09-08

Fetched 2026-09-08 23:01 UTC · 816483cf8f30597f

Torn from — the listing only (the original would not open)

“GPT-6 Astra from OpenAI is now generally available on Amazon Bedrock. It brings deeper reasoning and sharper judgment to your m…”

Published 2026-09-08

Fetched 2026-09-08 23:01 UTC · 61555b0baa6fc6fd

Torn from — the listing only (the original would not open)

“Last month, a Claude user noticed his account was consuming tokens even though he wasn't working. Anthropic has since warned us…”

They claim — On August 4, Grant De Swardt, an independent AI consultant in East Sussex, U.K., noticed something strange going on with his Claude Max 20x account.

Published 2026-09-08

Fetched 2026-09-08 23:01 UTC · 54678fe5fffcfda7

Torn from — the page itself

“Cognition's valuation multiple is higher than Cursor's was before selling to SpaceX.”

Published 2026-09-08

Fetched 2026-09-08 23:01 UTC · 01e544511c92e0f3

Torn from — the page itself