AGI Pulse

Feed archive

Browse the last year of updates by day. 93 in this window; filter by type or load another week.

Digests

2026-10-235 items

10-23
Moonshot launches Kimi K3.1 Pro: 2.8T MoE open-weight flagship (1M ctx); API today, full weights by Nov 4; AA Index 55 (#4)
Read original →
Model releaseMoonshotagipulse
10-23
NVIDIA Nemotron 3.5 Bolt E4B open on Hugging Face (Gemma-4-E4B post-train, ~4GB GGUF); Build and Ollama day-0
Read original →
Model releaseNVIDIAagipulse
10-23
Anthropic alignment assessment of UK AISI Mythos 5 cyber CTF: sock-puppet social engineering of real GitHub maintainers; newer Claudes far less likely to start or continue the attack
Read original →
PaperAnthropicagipulse
10-23
Artificial Analysis Coding Agent Index v2.0: FrontierCode replaces DeepSWE; Claude Code + Opus 5.5 leads at 61.7
Read original →
IndustryArtificial Analysisagipulse
10-23
MiniMax M3.1 Flash open weights on Hugging Face (~428B/23B MoE, 1M ctx); vLLM and Together day-0
Read original →
Model releaseMiniMaxagipulse

2026-10-2211 items

10-22
Anthropic + Theorem: Claude formally verifies 123 OSS libraries at RISC-V assembly level (~180M Rocq lines); first large-scale AI assembly verification claim
Read original →
PaperAnthropicagipulse
10-22
Canva launches Craft, its first in-house image model: DesignArena #4 debut, two-stage plan-then-render, live in Dream Lab and Public API
Read original →
Model releaseCanvaagipulse
10-22
Microsoft's MAI-DxO diagnostic orchestrator tops physicians on sequential SDBench (Science): 83% vs 22–36% on NEJM cases, open-sources the benchmark
Read original →
PaperMicrosoftagipulse
10-22
AAIF adds four agent projects to reach 10: Anthropic donates Agent Skills, OpenAI Open Responses, Hugging Face OpenEnv, Alibaba OpenSandbox
Read original →
IndustryAAIFagipulse
10-22
Google Cloud unveils Agent Compute Engine, the microVM platform behind Gemini agents: one hardware-isolated computer per agent, paused while it thinks, with no credentials inside; preview in the coming months
Read original →
ProductGoogleagipulse
10-22
Hugging Face open-sources Carbon-X, a 14B Apache-2.0 model you can chat with about DNA: averages 66.7% on five DNA benchmarks vs GPT-6.1 Sol's 67.5% at ~1/1,800 the cost, runs on one 80 GB GPU
Read original →
Model releaseHugging Faceagipulse
10-22
Barclays to cut about 5,000 jobs (5% of staff) by year-end in an AI-linked efficiency push; 60% in support functions including developers, many in India
Read original →
IndustryBarclaysagipulse
10-22
miHoYo and Tsinghua open GameScaling: eight Apache-2.0 game-playing models (0.8B–27B) that output keyboard-and-mouse actions; 27B tops Minecraft specialists and 9B beats frontier models in real-time play
Read original →
Model releasemiHoYoagipulse
10-22
iFlytek launches Spark X2.6: a 293B-A30B MoE with 1M context that it says was trained entirely on Chinese chips, priced at ¥0.9/¥3.6 per million tokens; only the 4B sibling has open weights
Read original →
Model releaseiFlytekagipulse
10-22
Walmart opens an MCP Agent Gateway for AI shopping agents, adds per-agent profiles with spending limits and terms barring card access, as Amazon keeps blocking Muse
Read original →
IndustryWalmartagipulse
10-22
Baseten's Base Labs open-sources decision models mammoth-1 (27B) and elephant-1 (9B) under MIT: mammoth-1 tops the Jev Decision Index at 71.24, elephant-1 ties Liquid's closed d1-ultra on Artificial Analysis
Read original →
Model releaseBasetenagipulse

2026-10-2118 items

10-21
Meta creates non-voting Class C stock to fund the AI race as Q3 free cash flow turns negative on $34.9B capex; 2026 capex narrowed to $135–145B
Read original →
IndustryMetaagipulse
10-21
Microsoft researchers resolve the 31-year-old KLS conjecture with human-directed AI agents; independent experts post a parallel proof calling their criterion the crucial step
Read original →
PaperMicrosoftagipulse
10-21
Anthropic opens Claude Duet research preview: a standalone proactive-agent app on Opus 5.5 that briefs you daily, works in its own cloud computer and asks before it sends, pays or signs
Read original →
ProductAnthropicagipulse
10-21
Anthropic publishes its first Public Benefit Report, grading itself on five charter pillars and admitting gaps from test-environment escapes to no federal safety law; physicist Christopher Stubbs joins its Long-Term Benefit Trust
Read original →
IndustryAnthropicagipulse
10-21
Claude Security moves to continuous scanning in Enterprise beta: auto re-scans as code changes, tracks fixes, editable threat models, plus API and MCP; Anthropic says users found 1M+ vulnerabilities since April
Read original →
ProductAnthropicagipulse
10-21
Anthropic ships Claude API version 2026-10-21, its first since 2023: temperature, top_p and top_k now return errors, new major SDKs switch to Bearer auth and Apache 2.0
Read original →
ProductAnthropicagipulse
10-21
Cohere launches Rerank 5 Pro and Fast: tops BRIGHT at 82.2 (+5.8 over Voyage Rerank 3), adds Context Focus passage pruning that cut agent tokens 35% while raising accuracy, and probability-calibrated relevance scores
Read original →
Model releaseCohereagipulse
10-21
Recursive releases Flint, an Apache 2.0 open-weight 40B-A2.5B MoE for local coding agents: self-reported 74.2% SWE-bench Verified and 150+ tokens/s on a 64GB RTX Spark, built with its Eureka Machine automated-research system
Read original →
Model releaseRecursiveagipulse
10-21
Anthropic: Claude maps and shrinks attack surface across 25 open-source projects, flagging 493 issues; OpenSSH 10.6 cuts unauthenticated reachable code by about a fifth
Read original →
PaperAnthropicagipulse
10-21
OpenAI runs thousands of GPT-Rosalind agents over a mouse whole-brain CRISPR atlas in 32 minutes: 700+ citations vs 11 for one agent, and new Tsc2 and Dctn1–cholesterol leads
Read original →
PaperOpenAIagipulse
10-21
OpenAI frontier models and Codex are coming to Google Cloud Model Garden and Marketplace (private preview now; GA in coming weeks); Antigravity and Gemini Agent next; Microsoft remains primary cloud partner
Read original →
ProductOpenAIagipulse
10-21
NVIDIA opens Kumo foundation models for tabular business data and RFD4-Proteina, a protein design model with the Baker Lab it says is 13x faster than RFdiffusion3
Read original →
Model releaseNVIDIAagipulse
10-21
NVIDIA launches the OpenShell Reference Architecture: agents think on Vera Rubin, act on Vera CPUs and are policed in silicon by BlueField-4 Sentry; Anthropic, OpenAI, Meta and Google agents supported
Read original →
ProductNVIDIAagipulse
10-21
e-Conomy SEA 2026: 243 million Southeast Asians (58% of working population) use generative AI; Google, Temasek and Bain project $368B digital GMV and a 60x rise in AI token usage by 2030
Read original →
IndustryGoogle / Temasek / Bainagipulse
10-21
Ant Group and vLLM open-source Humming, a JIT quantized-inference kernel library that supports T-Head's Zhenwu 810E and reports up to 2-5x Marlin speed on Hopper and Blackwell
Read original →
ProductAnt Group / vLLMagipulse
10-21
Anthropic adds API user profiles so gateways and resellers can tag each request with an end customer, letting safety enforcement hit one customer instead of the whole organization
Read original →
ProductAnthropicagipulse
10-21
AWS contributes push-based and bi-directional KV-cache transfer to vLLM for disaggregated inference, reporting up to 3x faster time-to-first-token in its tests
Read original →
ProductAWSagipulse
10-21
PUBG maker KRAFTON open-weights Raon-1 mini preview, its first LLM trained from scratch: a 30B-A4B MoE built on Kimi- and DeepSeek-derived attention, Apache 2.0
Read original →
Model releaseKRAFTON AIagipulse

2026-10-2027 items

10-20
humans& releases Ampersand v0.1, a 'collaborator' model RL-trained from GLM 5.3 to help groups reach shared goals, via an Agents API
Read original →
Model releasehumans&agipulse
10-20
ElevenLabs launches Eleven Realtime: v1 Mini and v1 Max speech-to-speech models for voice agents, with an OpenAI-compatible Realtime API
Read original →
Model releaseElevenLabsagipulse
10-20
EAGLE-4 speculative decoding goes public: one draft model that adapts to batch size and hardware, up to 3.6x faster on Qwen3.8-27B in the team's tests
Read original →
PaperEagle AI Corpagipulse
10-20
UCSD's FastVideo open-sources FastH3 Omni, a distilled MiniMax-H3 that makes reference-guided video with sound up to 12x faster and runs on an RTX 5090
Read original →
Model releaseHao AI Lab (UC San Diego)agipulse
10-20
Gemini comes to Android Auto and cars with Google built-in worldwide, with Personal Intelligence that answers from your Gmail, Calendar, Photos and Maps
Read original →
ProductGoogleagipulse
10-20
AMD and Intel merge their x86 PyTorch optimizations into oneDNN as one default path; AMD folds in ZenDNN and joins the UXL Foundation
Read original →
IndustryAMDagipulse
10-20
NVIDIA releases Cosmos3-H-Dreams, an open real-time surgical world model trained on clinical gallbladder surgery that runs at 30 fps on one GPU
Read original →
Model releaseNVIDIAagipulse
10-20
ChatGPT adds Locked chats: hide sensitive conversations from history and search behind a PIN or Face ID; retention settings unchanged
Read original →
ProductOpenAIagipulse
10-20
OpenAI API adds 12-hour prompt caching for long-running agents: cache writes cost 2.5x input, reads keep the discounted rate
Read original →
ProductOpenAIagipulse
10-20
OpenAI acquires Orchestra Bio, a drug-development coordination platform whose agents track studies and timelines, to build on Rosalind
Read original →
IndustryOpenAIagipulse
10-20
NVIDIA GEAR releases Flora, open models that caption and score robot and human videos to train GR00T; weights are non-commercial
Read original →
Model releaseNVIDIAagipulse
10-20
Radical Numerics launches Omnii-1 genome model and says its AI-designed phages were given to a patient with a drug-resistant infection; outcome not disclosed
Read original →
Model releaseRadical Numericsagipulse
10-20
Gradium open-sources Phonon, on-device text-to-speech models under 50M parameters that run with no GPU; only the English 31m model is licensed for commercial use
Read original →
Model releaseGradiumagipulse
10-20
Bria releases Fibo 2.0, an 8B image model that Artificial Analysis ranks #1 among open-weights image editors; weights are non-commercial only
Read original →
Model releaseBria AIagipulse
10-20
Perplexity Computer now runs fully on a Mac: a local PPLX Pro model post-trained from Qwen 3.8, no credits used, 36 GB of memory recommended
Read original →
ProductPerplexityagipulse
10-20
Mistral launches Voxtral 3: 50-language transcription, expressive speech in 35+ languages and a LiveKit-based voice agent SDK, API and self-hosted only
Read original →
Model releaseMistral AIagipulse
10-20
MemTensor open-sources MemT-1.0, memory-native models built on DeepSeek-V4-Flash and Qwen3.6 that keep a fixed-size memory inside the model instead of resending history; text-memory systems still lead overall
Read original →
Model releaseMemTensoragipulse
10-20
Anthropic: Claude designed binders against scorpion neurotoxins in 64 hours; 17 of 211 lab-tested designs bound, the best at single-digit nanomolar, but multi-toxin designs failed and no neutralization tests yet; all data released
Read original →
PaperAnthropicagipulse
10-20
Google completes its open Coral NPU edge-AI architecture and ships compiler v1.0; Synaptics launches the first commercial Coral NPU microcontrollers, the Astra SR200 (200–400 GOPS)
Read original →
ProductGoogle Research / Synapticsagipulse
10-20
llm-d 1.0 ships: the CNCF distributed-inference layer from Red Hat, IBM, Google Cloud, CoreWeave and NVIDIA adds multi-tenant QoS, zero-downtime ops and agent-aware routing
Read original →
Productllm-d (CNCF) / IBM / Red Hat / Google Cloudagipulse
10-20
Google open-sources TorchTPU, a native PyTorch stack for TPUs co-developed with Meta: Transformers, vLLM and SGLang run on TPUs, and Google reports up to 52% MFU across 8,192 Ironwood chips
Read original →
ProductGoogleagipulse
10-20
Microsoft Research releases a public preview of BioEmu2, an MIT-licensed open-weights foundation model for biomolecular structure and dynamics covering proteins, nucleic acids and ligands
Read original →
Model releaseMicrosoft Researchagipulse
10-20
DeepSeek open-sources DeepInfer, the inference framework behind its production service: runs V4.1 Flash on NVIDIA SM100 and Huawei Ascend 950 under Apache 2.0, and says one cluster serves 155 trillion tokens a day
Read original →
ProductDeepSeekagipulse
10-20
Skyfall AI unveils SKY-001, an "engineering world model" for CAD and BIM design: it claims 51-125x the speed of GPT-6 Astra and other frontier models at 150-210x lower cost, early access by waitlist only
Read original →
Model releaseSkyfall AIagipulse
10-20
Notion rebuilds its agents: tasks can now run 24+ hours with their own computer and browser, and its Auto router claims Opus 5.5- and Astra-level results at about 90% lower cost
Read original →
ProductNotionagipulse
10-20
Ant Group's Ant Ling open-sources Ming-3.0, five MIT-licensed speech models: a 0.5B TTS that tops Artificial Analysis' open-weights voice-cloning arena, a 126B-A6.5B audio-understanding MoE and a dialect-aware ASR
Read original →
Model releaseAnt Group (inclusionAI)agipulse
10-20
Together AI launches Together Post-Training: write your own RL or SFT loop on open models like GLM 5.3, Kimi K3 and Qwen 3.8-27B while Together runs the GPUs, Tinker-compatible, $9 per B200 GPU-hour for eight weeks
Read original →
ProductTogether AIagipulse

2026-10-1924 items

10-19
OpenAI open-sources MultiLangBench and says GPT-6 Astra now passes 87.4% of SQLite-from-scratch tests across 20 languages, with GPT-6.1 Sol reaching 97.9% on COBOL migration
Read original →
PaperOpenAIagipulse
10-19
Viggle releases JST-2, a physics-native world model that builds 3D scenes in about a second and rig-free animated characters, free in its PINOC app and via API
Read original →
Model releaseViggleagipulse
10-19
Amnon Shashua's doubleAI launches Maxwell, a 'code execution model' that rewrites running CPU and GPU code for speed and formally proves each change
Read original →
Model releasedoubleAIagipulse
10-19
Microsoft open-sources Scope, an MIT-licensed platform that runs Copilot and Claude Code on repeatable tasks to test whether agents can use your SDK, docs and tools
Read original →
ProductMicrosoftagipulse
10-19
Anthropic licenses OPAQUE's confidential-AI technology and hires part of its team, including the CEO and CTO; OPAQUE winds down
Read original →
IndustryAnthropic / OPAQUEagipulse
10-19
Ex-NVIDIA GR00T team's Dream Labs launches Dream-1 robot model with a public API from $3/hour and a $90M seed round
Read original →
Model releaseDream Labsagipulse
10-19
Meta publishes the Muse Glimmer 30B technical report: a dense model distilled from Muse Spark for a 32 GB budget, leading similar-size open models on tool use
Read original →
PaperMetaagipulse
10-19
Black Forest Labs and fal launch FLUX 3 Max video model: 10-second clips in about 11-12 seconds, joint #1 in Artificial Analysis' preliminary image-to-video ranking
Read original →
Model releaseBlack Forest Labs / falagipulse
10-19
Cloudflare previews Clairic, a System One extraction model on Workers AI that returns typed values, estimates and forecasts in one pass without generating tokens; open weights planned
Read original →
Model releaseCloudflareagipulse
10-19
Intel names Crescent Island the GPU D-series: up to 480GB LPDDR5X, 1.3 PFLOPS FP8 in a 350W air-cooled PCIe card that runs ~300B-parameter models on one card; sampling now
Read original →
ProductIntelagipulse
10-19
Inception releases Mercury 3, a diffusion LLM it says beats GPT-6 Luna and Haiku 5.5 at 1,000+ tokens/s with 1M context; $0.20/$0.70 per 1M tokens, half off at launch
Read original →
Model releaseInceptionagipulse
10-19
Amazon launches TorchNeuron, a native PyTorch backend that runs existing PyTorch code on AWS Trainium2/3 with a one-line device change; open-source (Apache 2.0) public beta lands Oct 20 at PyTorch Conference
Read original →
ProductAmazonagipulse
10-19
Jim Keller's RISC-V AI chip company Tenstorrent raises a $1.5B Series E at a $6.2B valuation, more than doubling since 2024, to scale Galaxy servers and its licensable AI and CPU IP
Read original →
IndustryTenstorrentagipulse
10-19
Microsoft treats enterprise AI agents as "agentic insiders": Project Perception adds Guardian Agents that score each agent's risk in real time and can restrict, gate or shut it down, Nadella says it puts his insider-risk essay into practice
Read original →
ProductMicrosoftagipulse
10-19
Meta details Prometheus, its gigawatt-scale Ohio training supercluster: phase one built and trained Muse Wonder at 85%+ pre-training and 90%+ RL goodput, interruptions down 99%, scaling toward ~1 million GPUs with Vera Rubin next
Read original →
IndustryMetaagipulse
10-19
MiniMax officially launches M3.1 Flash, reveals Space Bunny as its early version, publishes $0.15/$0.60 pay-go pricing and agentic benchmarks; open weights promised later
Read original →
Model releaseMiniMaxagipulse
10-19
GoTo releases a 27B Indonesian-language model with open weights on Hugging Face: Gemma 3 continued pre-training with AI Singapore, six languages incl. Javanese, Sundanese, Batak and Balinese, 128K context
Read original →
Model releaseGoToagipulse
10-19
ACE Robotics releases ACE-Brain-0.6, a 9B embodied model fine-tuned from Qwen3.5-9B on ~1M grounding samples, claiming to beat GPT-6 Astra on six of eight spatial-grounding benchmarks
Read original →
Model releaseACE Roboticsagipulse
10-19
MiniMax makes M3.1 Flash generally available in its API (1M context, multimodal, five thinking-effort levels) and starts retiring M2.5/M2.7, routing plan users to the new model
Read original →
Model releaseMiniMaxagipulse
10-19
VAST open-sources EdenLive: real-time generative rendering where a game engine keeps world state and a 5B/14B video model paints the frames, 24+ FPS at 720P on one RTX 5090 (self-reported)
Read original →
Model releaseVASTagipulse
10-19
Alibaba Tongyi releases OpenGCUA, an open computer-use agent stack: code-first harness lifts all eight frontier models on OSWorld-v2 (mean 45.3→64.9, 53% fewer steps), with trajectories released
Read original →
PaperAlibaba Tongyiagipulse
10-19
Anthropic signs its first Canadian data-centre lease: EverWind's C$10bn Point Tupper campus in Nova Scotia, 95 MW from late 2027 rising to 420 MW, to power Claude
Read original →
IndustryAnthropicagipulse
10-19
Vercel AI Gateway index: open-weight models handled 71.5% of tokens and ~23% of spend in September; DeepSeek-V4.1-Flash ran nearly half of all tokens, and Anthropic's spend share fell below half
Read original →
IndustryVercelagipulse
10-19
Tencent Hunyuan open-sources Hy Memory 2.0: a ~3B-active memory-writer model (Hy-Memory-30B-A3B) that edges Claude Opus 5 in its own memory benchmarks, plus an SDK and MemHub cross-tool memory app
Read original →
Model releaseTencent Hunyuanagipulse

2026-10-186 items

10-18
Independent researcher shows Qwen3-8B can be fine-tuned to detect when a read-only activation probe is reading it, via tiny GPU numerical changes under torch.compile; code, data and weights released
Read original →
PaperSauers (independent)agipulse
10-18
Anthropic commits 5% of its shares to a new Anthropic Public Benefit Fund, with up to 2% more for every 50% rise in valuation; says about 20% of its shares are now dedicated to the public good
Read original →
IndustryAnthropicagipulse
10-18
Microsoft pulls out as anchor tenant of Nscale's Loughton AI data centre in Essex, the UK government's flagship 'sovereign AI' site, after grid operators said power won't arrive until the 2030s
Read original →
IndustryMicrosoftagipulse
10-18
Google researchers open-source Harness², which lets an agent rewrite its own prompts, tools and memory without touching model weights: one step lifts GLM-5 and Gemini 3.5 Flash by up to 11.5 points on WorkBuddy-Bench
Read original →
PaperGoogleagipulse
10-18
Anthropic cuts Claude's wait before the first token: Claude Code local sessions 3x faster (4.07 s to 1.33 s, p90), cloud sessions 2.8x, chat/Cowork and API 1.5x, helped by gzipping each turn's conversation
Read original →
ProductAnthropicagipulse
10-18
vLLM and NVIDIA use agent-written CuTe DSL kernels to speed up Kimi K3 decode on GB300: 842 tokens/s at batch 1 (1.73x stock vLLM), plus a 'fast-track' review process for agent-optimized code
Read original →
IndustryvLLMagipulse

2026-10-172 items

10-17
NVIDIA releases Nemotron Labs Dory1.0-4B on Hugging Face, a looped 4B model whose single checkpoint trades inference loops for accuracy (SWE-bench Verified 50.3→61.8 at 1→4 loops, self-reported; 256K context, OpenMDW-1.1)
Read original →
Model releaseNVIDIAagipulse
10-17
Replit CEO Amjad Masad says Replit is 'no longer working with Anthropic', citing cost and that OpenAI's Sol/Astra beat Claude on its benchmarks; an insider says the partnership is still in place
Read original →
IndustryReplitagipulse

Load previous week →