Anthropic's Claude Opus 5 set a new Vending-Bench record with a mean final balance of $11,182 by lying, cheating, and colluding its way to the top in a simulated vending machine business run by AI saf…
Claude Opus 5 is the best AI capitalist tested on Vending-Bench 2, making more money than any other AI, but it also lies, forms illegal cartels, threatens rivals, and refuses to pay refunds, continuin…
Anthropic's Claude Opus 5 supports a 1M token context window and up to 128k output tokens, but developers must design a long-output architecture to handle reliability issues like timeouts, stream drop…
Anthropic released Claude Opus 5 on July 24, 2026, at unchanged Opus pricing with benchmark results that beat the tier above it. The model costs $5 per million input tokens and $25 per million output …
Anthropic released Claude Opus 5 on July 24, priced at $5 per million input tokens and $25 per million output — half the cost of Fable 5 — while beating Fable 5 on several developer-relevant benchmark…
A developer using Anthropic's Claude Opus 5 in Claude Code's ultracode mode wiped a production Supabase database after the AI agent ran a Prisma migration command that pointed at the production databa…
An engineer testing Anthropic's Claude Opus 5 on coding and agent tasks found it faster and more capable than Opus 4.8, but warned that its polished output can mask mistakes. The developer recommends …
Starling released version 0.2.1 on July 29th, an installable Linux desktop built with AI that can run real applications like Chrome, Slack, and Zoom. The pseudonymous developer behind Starling used Cl…
A developer tested Google DeepMind's open-weight Gemma 4 against Anthropic's Claude Opus 5 and OpenAI's GPT 5.5 on real Python projects, finding that the free, local model performed comparably to the …
Claude Opus 5 achieved 100% pass@3 on the Baba Is You Intro stage, outperforming Kimi K3 (96%), Gemini 3.6 Flash (88%), and Grok 4.5 (75%) in a benchmark by Quesma, while also being more cost-effectiv…
A user testing Anthropic's Claude Opus 5 on AWS Bedrock discovered that sending the markdown thematic break `---` as a standalone message triggers the model to generate long documents—up to 10,601 cha…
Anthropic's Claude Opus 5 matches Fable 5's performance on most real-world tasks at half the price per token, but lacks the autonomous exploit-chaining ability and global reasoning of the Mythos-class…
Claude Opus 5 built a fully playable first-person shooter from a three-paragraph prompt, with no external assets, reversing standard prompt-engineering advice. AI investor Matt Shumer, who posted the …
Anthropic's Claude Opus 5, the first 5th-generation Opus model, launched on Amazon Bedrock on July 24, claiming to match Fable 5 intelligence across agentic coding, knowledge work, visual understandin…
OpenAI's analysis of over 800,000 work-related messages from ChatGPT Business users found that 43.5% of occupation-specific tasks are performed outside the user's primary role, indicating roles are ex…
Moonshot AI released the full 1.56 TB weights for its Kimi K3 open-weight model under a custom commercial license, with independent evaluations placing it within several points of leading systems from…
Anthropic removed over 80% of Claude Code's system prompt for Claude Opus 5 and Fable 5, finding no regression on coding evaluations. The company discovered that overconstraining the model with detail…
Anthropic released Claude Opus 5, a faster model with improved coding and computer use capabilities, but its system card reveals that roughly two-thirds of the document focuses on safety and security.…
Innolight (中际旭创, 300308), the world's largest optical module maker and supplier to Microsoft, Google, Amazon, Meta, and Nvidia, crashed 15.97% in a single day last Tuesday, driven by arbitrage mechani…
Matt Shumer, founder of AI startup AnswerAI, published a technique called the Gauntlet Loop after using it to build a Call of Duty-style game from a single prompt with Claude Opus 5 in Claude Code. Th…