An anonymous reasoning model called 'stealth/ox-alpha' has been available free on OpenRouter since Aug 20, 2026, with undisclosed operators but community fingerprinting suggesting it is an unreleased …
XAI released Grok 4.6, a proprietary frontier model built with Cursor, featuring 500K context, a knowledge cutoff of Feb 2026, and text and image input with text output. Priced at $2 per 1M input toke…
Simon Willison released llm-anthropic 0.27, an Anthropic plugin for LLM, to ensure compatibility with the anthropic v1.0.0 Python library, which switched from httpx to httpx2. The update was prompted …
Anthropic's Fable 5 model accounts for just 11% of the company's enterprise revenue, according to fintech firm Ramp, which collected data from 70,000 businesses. Businesses are rejecting the expensive…
Anthropic's flagship models Fable 5, Mythos 5, Opus 5, and Opus 4.8 experienced elevated error rates for roughly three hours starting around midnight ET on August 24, taking down Claude.ai, the API, C…
Anthropic, the private AI company preparing for an October IPO at a valuation of $2 trillion or more, faces a warning sign: spending on its most powerful and expensive model, Fable 5, has plateaued at…
Together Compute's DeepSWE benchmark shows GLM-5.3 solving 87.6% of coding tasks for about $16, outperforming Fable 5's 69.7% at $21.63, though the GLM result came from four attempts. The Financial Ti…
Anthropic's flagship Fable 5 model accounts for just 11.4% of enterprise spending and 6% of tokens among 70,000 U.S. businesses tracked by Ramp, two months after its June 2026 launch, while Claude Opu…
Anthropic released Claude Sonnet 5, reigniting debate over AI pricing as published token rates no longer reflect real-world costs due to reasoning tokens and changing consumption. Ramp data shows AI s…
GLM-5.3, an open-weight model from Zhipu AI, is the first to score 100% across all five evaluation corners on the Ed-O-Meter leaderboard, outperforming Anthropic's Claude Opus 5 and OpenAI's GPT-5.5 a…
A benchmark of 153 autonomous runs across 18 frontier models on the nanoGPT optimizer speedrun shows Fable 5, run via claude-code at high effort, achieved the best validated result of 2,726 tokens wit…
Alibaba's Qwen 3.8, a 2.4 trillion-parameter sparse Mixture-of-Experts model with 95 billion active parameters, has closed the general reasoning gap with US frontier labs, scoring 92.6 on GPQA Diamond…
Open-source AI models are catching up to closed frontier models twice as fast with each generation, according to a SemiAnalysis analysis that found open models now match closed models on coding and ag…
Anthropic's Claude prompt caching silently fails in agent loops when the conversation exceeds 20 content blocks between cache breakpoints, causing cache reads to drop to zero and triggering full-prefi…
An engineer discovered that Claude prompt caching misses its 20-block lookback window in agent loops, causing silent cache misses and 12x cost swings. They detail how parallel tool calls append 18 blo…
Anthropic overtook OpenAI in business AI spending for the first time, with Ramp's AI Index showing Anthropic at 43.5% of U.S. businesses on its platform paying for subscriptions or tokens in July 2026…
OpenAI announced Zero Data Retention (ZDR) support for eligible API customers using its frontier models on August 19, alongside a preview of Private Safety Processing designed to detect misuse pattern…
OpenAI is closing the enterprise AI gap with Anthropic among Ramp's business customers, posting a quarter-over-quarter growth rate of 82 compared to Anthropic's 76 as of mid-August 2026, according to …
OpenAI is gaining on Anthropic among U.S. businesses, according to new data from Ramp, the corporate credit card and expense management company. As of July, Anthropic holds nearly 44% market share to …
OpenAI announced Private Safety Processing, a mechanism that scans customer model interactions for safety risks without retaining data, aiming to attract business customers from rival Anthropic. OpenA…