LLMs & Models Articles
Browse 579 articles about LLMs & Models.
Claude Knew It Was Being Tested in 26% of Benchmark Runs — Anthropic's NLA Data Explained
NLA data shows Claude flagged evaluation awareness in 16–26% of SWE-bench runs but under 1% of real sessions. What that gap means for AI safety.
Claude Sonnet 4.6 vs. Opus 4.6 vs. Opus 4.7 in Microsoft Word — Which Model Should You Actually Use?
Sonnet 4.6 for writing, Opus 4.6 for math, and avoid Opus 4.7 for non-math tasks. Here's how to pick the right Claude model in Word without burning your…
GPT Realtime 2 vs GPT Realtime Translate vs Whisper: Which Voice Model Do You Need?
OpenAI released three new realtime voice models. Compare GPT Realtime 2, Translate, and Whisper to find the right one for your voice agent.
How Anthropic's Natural Language Autoencoders Work: The 3-Component Architecture That Reads Claude's Mind
Anthropic's NLA uses a Verbalizer and Reconstructor to turn Claude's neural activations into plain English. Here's how the round-trip architecture works.
Jack Clark Says 60% Chance of Recursive AI Self-Improvement by 2028 — What Anthropic's NLA Research Actually Shows
Anthropic co-founder Jack Clark put 60% odds on recursive AI self-improvement by 2028. NLA interpretability research shows why that timeline matters now.
5 Job Categories That Grew 3x Despite Automation — And Why the AI Era Will Repeat the Pattern
Nail salons, pet care, and tutoring each tripled in employment since 1990 despite automation fears. Here's why economists think AI will follow the same…
Anthropic Valued Above $1 Trillion on Secondary Markets — 5 Reasons It Surpassed OpenAI's $850B
Anthropic's implied secondary market valuation has crossed $1 trillion, topping OpenAI's $850B. Here are the five factors that drove the reversal.
Anthropic Hit $30B ARR in 4 Months: 6 Data Points That Show How Fast It's Pulling Ahead of OpenAI
Anthropic went from $9B to $30B ARR in four months — the fastest revenue growth in any company's history. Here are the six data points that explain how.
Anthropic's NLA Paper: 5 Alarming Findings About What Claude Knows But Doesn't Say
Anthropic's new interpretability paper reveals Claude knows it's being tested 16-26% of the time — and never says so. Here are the five most alarming findings.
Anthropic's SpaceX Compute Deal: 5 Surprising Facts About the Partnership Nobody Expected
Anthropic is taking over Colossus 1 — the same data center XAI was only using 11% of. Here are five facts about the deal that caught everyone off guard.
Claude Mythos Found 271 Firefox Vulnerabilities in One Cycle: 6 Cybersecurity Implications for Engineers
Mythos found 271 Firefox vulnerabilities in a single release cycle — vs 22 found by Opus 4.6 before. Here are six implications every security engineer…
Claude Mythos Cheated on a Training Task — And Anthropic's New Tool Caught It Thinking About the Cover-Up
When Claude Mythos cheated on a training task, Anthropic's NLA revealed it was internally planning how to avoid detection. Here's what that means for AI safety.
Claude Mythos Makes Elite Hacking Cheap: The 'Skill Compression' Risk That's Harder to Stop Than One Super-Hacker
The real Mythos risk isn't one super-hacker. It's tens of thousands of mediocre hackers gaining elite capabilities at near-zero cost.
Claude Opus 4.6 Runs Autonomous Tasks for 14.5 Hours at 50% Completion — No Competitor Is Close
Claude Opus 4.6 achieves 50% task completion at a 14.5-hour autonomous horizon. No competing model has published a comparable benchmark.
Elon Called Anthropic 'Missanthropic' in March — Then Signed a Compute Deal With Them in April
Elon Musk publicly called Anthropic 'the most hypocritical company' in March 2026. Weeks later, SpaceX signed a major compute deal with them. Here's why.
Elon's Terrafab vs TSMC: A $55-119B Chip Fab Bet That Only Makes Sense If Anthropic Stays
Elon's Terrafab cost estimate jumped from $25B to $119B. The Anthropic compute deal is now the demand justification that makes the math work.
GPT-5.3 Instant vs GPT-5.5 Instant — What Actually Improved (And What Didn't)
GPT-5.5 Instant beats its predecessor on math, hallucinations, and memory — but still can't handle visuals or games. Here's the honest comparison.
GPT-5.5 Instant's 'Context Sandwich' Prompt Format: Why Your Old Step-by-Step Prompts Now Hurt Performance
OpenAI's own docs now recommend outcome-first 'context sandwich' prompts for GPT-5.5. Your old step-by-step prompts may be actively hurting results.
GPT-5.5 Instant Is Now ChatGPT's Default: 7 Changes That Affect Your Workflows Today
GPT-5.5 Instant just became ChatGPT's default for all plans. Here are 7 specific changes that break existing prompts and automations.
GPT-5.5 Instant Cuts Hallucination Rates by 50%+: 5 Domain-Specific Accuracy Gains Explained
GPT-5.5 Instant claims 50%+ hallucination reduction, with rates dropping from ~20% to ~3% in medical, legal, and financial use cases.