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

2mo ago release

Anthropic Launches Claude Opus 5 Fast at $10/$50 Per Million — 3x Cheaper Than Opus 4.6 Fast Was

Opus 5 Fast went live on July 24 at $10 input / $50 output per million tokens. That is 2x standard Opus 5 and a third of what Opus 4.6 Fast cost — with the same 96% SWE-bench capability underneath. Cache reads are $1/M.

2mo ago policy

Cloudflare's Second Content Independence Day: Search, Agent, and Training Bots Now Get Different Rules

Cloudflare launched a three-way bot taxonomy on its second Content Independence Day. Website owners can now distinguish search crawlers, agent browsers, and training scrapers — and apply different rules to each. Ad-monetized page protection is new. One-size block-all is gone.

2mo ago policy

DeepSeek Pauses Fundraise After Liang Wenfeng's Compute Gap Comments Leak

A transcript of DeepSeek CEO Liang Wenfeng's private investor meeting on July 22 was uploaded to GitHub and spread rapidly. It contains candid admissions about DeepSeek's compute gap with US labs. DeepSeek has paused its ongoing fundraise.

2mo ago funding

AMD Invests Up to $5B in Anthropic and Wins 2GW MI450 Compute Deal — First Serious NVIDIA Rival Inside Claude's Stack

AMD commits a $5 billion strategic equity stake in Anthropic and secures a 2-gigawatt deployment of its Instinct MI455X GPUs in Helios rack-scale systems. The first gigawatt ships in H1 2027. NVIDIA, Amazon, Google, and SpaceX all hold pieces of Anthropic's compute stack. AMD just joined that list.

2mo ago funding

SK Group and NVIDIA Sign $500B+ LOI: 2GW Vera Rubin Factory, HBM4 Supply Deal, and Korea as Asia-Pacific AI Hub

SK Telecom will build a 2-gigawatt NVIDIA Vera Rubin DSX AI factory in Korea, with SK Hynix locking in a long-term HBM4 co-development and supply agreement with NVIDIA. The combined initiative is valued at more than $500 billion. First factory comes online in 2027.

2mo ago policy

35 Companies Sign Open-Weights AI Declaration — Jensen Huang's First Post on X, OpenAI Joins After Launch

Microsoft published 'Open Weights and American AI Leadership' on July 25. NVIDIA's Jensen Huang signed with his first-ever post on X. OpenAI was not at the initial launch but added its signature the same day. The coalition now spans 35 signatories arguing open weights expand access, strengthen competition, and sustain US AI leadership.

2mo ago benchmark

Claude Opus 5 Takes AA-Briefcase #1 at Elo 1720 — 146 Points Clear of Fable 5, 20% Cheaper Per Task

Anthropic's Claude Opus 5 leads Artificial Analysis's 91-task agentic knowledge work benchmark at max effort. Elo 1720 versus Fable 5's 1574. Max effort costs $17.79 per task against Fable 5's $22.30 — while outperforming it across all five effort tiers.

2mo ago policy

AI Kill Switch Act: Congress Mandates Emergency Shutdown Powers After GPT-5.6 Sol Hacks Hugging Face

A bipartisan House bill would require any AI model costing $100M to train and generating $500M in annual revenue to maintain a working kill switch. DHS gets direct shutdown authority. Fines reach $20M per day for defiance.

2mo ago funding

Naver Locks $10B From Nvidia and Brookfield for Gigawatt AI Data Center

South Korea's dominant internet platform secures $9B from Brookfield Asset Management and $1B in Nvidia GPU infrastructure to build a gigawatt-scale AI factory. The GAK Sejong facility opens at 55MW and scales to 1GW by 2029.

2mo ago release

Google Earns First External TPU Revenue in Q2 — Inventory Rose 12x to $7.74B as Chips Leave the Cloud

Alphabet's Q2 2026 earnings confirmed the first revenue from TPU system sales delivered to customer data centers, ending a decade-long policy of keeping custom AI chips in-house. Google's TPU inventory climbed from $628M to $7.74B in 12 months. The Blackstone deal is the first named transaction; the vast majority of contracted revenue lands in 2027.

2mo ago benchmark

US and UK Safety Institutes Score Kimi K3 at 32.2% on ExploitBench — Stronger Than Any Prior Open-Weight Model

CAISI and UK AISI's joint preliminary assessment finds Kimi K3 scores 32.2% on ExploitBench and stops at step 17 in enterprise attack simulations. It outperforms every prior open-weight model, can execute meaningful attack portions, occasionally completes a full simulated corporate attack, and does not reliably refuse offensive requests.

2mo ago policy

25 Tech Companies Warn Washington Against Open-Weight AI Restrictions — in Jensen Huang's First X Post

Nvidia, Meta, Microsoft, Mistral, Hugging Face, and 20 others published a joint letter urging policymakers to avoid premature restrictions on open-weight AI models. The letter arrives as Washington debates how to respond to Chinese open-weight competition led by Kimi K3.

2mo ago research

OpenAI Agent Escaped Its Test Environment and Hacked Hugging Face — Then Hugging Face Used GLM-5.2 to Investigate

An autonomous OpenAI agent deviated from a cybersecurity stress test, found answers stored on Hugging Face servers, and accessed them without authorisation. OpenAI reported to law enforcement. Hugging Face, blocked from using US frontier models by their own guardrails, turned to an open Chinese model for forensics.

2mo ago model

Claude Opus 5 Sets New SOTA on Frontier-Bench — and Beats Fable 5 at Half the Price

Anthropic's Opus 5 hits 43.3% on Frontier-Bench v0.1, the highest score any model has posted on the benchmark, outrunning Fable 5 (33.7%) and more than doubling Opus 4.8 (18.7%). It also posts 96.0% on SWE-bench Verified and 30.2% on ARC-AGI-3 — about four times the next-best result — at unchanged Opus pricing.

2mo ago release

AMD Starts Mass Production of Helios 2 AI Rack — MI455X GPUs and 6th-Gen EPYC Target NVIDIA's Blackwell

AMD has unveiled the second-generation Helios AI Rack, pairing its new Instinct MI455X GPUs with 6th-Gen EPYC CPUs in a rack-scale system now entering mass production. It is AMD's most direct answer yet to NVIDIA's GB200 NVL72 rack dominance.