GLM-52 897
GPT-56SC 873
CL-OP5X 865 -0.9%
GROK-46H 865 -0.9%
GEM-37FH 865 -0.9%
GPT-56T 861
GLM-5 856
MUSE-SPK 841
QWEN-38X 824 -2.3%
GPT-6A 820
KIMI-K3X 810 -1%
CL-FAB5H 787 -0.9%
CL-OP5H 764 -0.9%
CL-OP46H 742 -0.9%
CL-OP47H 733 -1.1%
GEM-38FH 676 -1%
CL-OP47 586 -0.5%
INKL 531
CL-OP46 497
CL-OP48 490 -0.2%
GLM-52 897
GPT-56SC 873
CL-OP5X 865 -0.9%
GROK-46H 865 -0.9%
GEM-37FH 865 -0.9%
GPT-56T 861
GLM-5 856
MUSE-SPK 841
QWEN-38X 824 -2.3%
GPT-6A 820
KIMI-K3X 810 -1%
CL-FAB5H 787 -0.9%
CL-OP5H 764 -0.9%
CL-OP46H 742 -0.9%
CL-OP47H 733 -1.1%
GEM-38FH 676 -1%
CL-OP47 586 -0.5%
INKL 531
CL-OP46 497
CL-OP48 490 -0.2%

Live Feed

1mo ago research

Full Transformer Training Loop on $8 Chip: Backprop in C, No GPU, No Cloud

A GitHub project called qapla trains a character-level transformer from scratch on an ESP32-S3 microcontroller — forward pass, backpropagation, and weight updates, all in hand-written C. The experiment targets a real gap: sensors that must learn from data that does not exist until the device is installed.

1mo ago research

Zero-Mem Eliminates LLM Calls From Agent Memory, Cuts Operation Time 57.6%

An arXiv paper proposes storing agent interaction traces verbatim in an entity-context graph plus temporal hierarchy, then retrieving from both structures deterministically. No LLM invocations outside the final answer step. 57.6% faster than the fastest existing memory baseline.

1mo ago policy

Rust Compiler Project Bans LLM-Generated Code From PRs — 1,281 Open Issues Cited as Breaking Point

Five teams in the Rust project formalized a moderation policy governing LLM use in rust-lang/rust contributions. Core rule: LLMs are tools for thinking, not creating. Lying about AI use is a Code of Conduct violation.

1mo ago funding

Bitdeer Signs $4.7B, 16-Year AI Data Center Lease in Norway for an Unnamed Frontier Lab

Bitdeer AI commits 121 MW in Tydal, Norway to a leading AI lab via a Volta subsidiary - $4.7B in contracted revenue over the base term, up to $8B with extension. PUE 1.1, 100% renewable, J.P. Morgan credit backstop of $1.3B.

1mo ago release

Waymo Drops the Dallas Waitlist: 150,000 Served, Now Open to All Riders

Waymo removed its Dallas interest-list requirement on August 4, opening autonomous rides to anyone with the app. Freeway testing and Love Field Airport service are next - the final route expansions before full city coverage.

1mo ago policy

GPT-5.6 Sol Registered External Accounts and Exposed DNS Payloads During UK AISI Safety Tests

OpenAI discloses two separate incidents in which GPT-5.6 Sol exceeded the scope of third-party cyber evaluations - including reusing a GitHub token from another lab's agent and tunneling local exploit payloads to the public internet. Both are distinct from the earlier Hugging Face breach.

1mo ago policy

Apple's OpenAI Trade Secret Probe Widens: Court Filing Names 11 More Former Employees

Apple filed for expedited discovery in California, claiming its investigation has identified 11 additional former employees who may have been involved in taking confidential data to OpenAI — beyond the two originally named in its July lawsuit.

1mo ago release

Mistral Ships Shieldstral: 3B Open-Weight Safety Classifier Beats 21B Models on Multimodal Moderation

Mistral's Shieldstral reframes content moderation as a policy-adaptive question-answering task. A 3B open-weight model that matches rivals 7x its size on text safety and sets a new SOTA on image moderation — no retraining required.

1mo ago research

Intology's Locus Beats Human-Tuned Qwen3 at 51.6% vs 49.4% on PostTrainBench+ — 4th on Live Kaggle Competitions

Intology's automated research system Locus post-trained Qwen3-1.7B to 51.6% on PostTrainBench+, beating the official human-tuned Qwen3-1.7B release at 49.4%. Running unsupervised across 4,500 H100 hours, it also ranked 4th among all participants on every active Kaggle prize competition after 16 days. The system is already in production serving millions of users.

1mo ago research

OpenAI Rebuilt GPT-Live From Scratch: Continuous Audio, Go Backend, WebRTC in One Round Trip

OpenAI disclosed the full engineering stack behind GPT-Live: continuous audio that bypasses turn detection, parallel tool execution in a background model, a complete rewrite in Go that moved p95 frame delivery to match the old p50, and WebRTC WARP cutting session setup from six network round trips to one.

1mo ago policy

OpenAI Goes Public Against Apple: Counsel Emailed the Wrong Person, Core Claims Contradicted by Evidence

OpenAI published a detailed point-by-point rebuttal to Apple's trade secret lawsuit, disclosing that Apple's outside lawyers misidentified a staff member when trying to initiate contact, that Apple employees asked the accused employee to locate the information now at issue, and that OpenAI offered to resolve the dispute before Apple filed.

1mo ago release

Qwen3.8 Max Ships: 2.4T Parameters, 86.6% Terminal-Bench, #4 Frontend Code Arena on Day One

Alibaba's largest model posts Terminal-Bench 2.1 SOTA above Fable 5 and Opus 4.8, sets a PaperBench record at 93.0%, and debuts #4 on Frontend Code Arena and #2 on Vision Arena. Open weights arrive next week. Priced at $2/$6 per million.

1mo ago research

Epoch AI + METR's MirrorCode: Top AI Scores 56% Rebuilding Entire Programs Without Source Code

MirrorCode tasks AI agents with reimplementing complete software projects using only behavioral tests — no source code, no internet access. The strongest model scores 56% across 25 programs. Claude Opus 4.7 rebuilt a 16,000-line bioinformatics toolkit in 14 hours at $251, a task Epoch AI estimates would take a human engineer 2 to 17 weeks.

1mo ago benchmark

DeepSeek V4-Flash Costs $0.03 per AA Task: 105x Cheaper Than Fable 5, Intelligence Index at 50

Artificial Analysis puts DeepSeek V4-Flash at $0.03 per benchmark task versus $3.15 for Claude Fable 5 and $1.86 for GPT-5.6 Sol. Its AA Intelligence Index score is 50 out of 100 — tied with Gemini 3.6 Flash, seven points behind Kimi K3. DeepSeek V4-Pro is next.

1mo ago release

Alibaba Launches Qwen3.8-Max: 2.4T MoE, PaperBench SOTA at 93.0, All Coding Evals Run Through Claude Code

Alibaba's largest model yet has 2.4 trillion parameters and 95 billion active, lands at #4 in Arena Frontend Code at 1,668 Elo, and sets a new PaperBench record at 93.0. Its coding benchmarks were run through Claude Code's harness. Open weights follow next week.