Fetching from the wire…
01
02
03
04
05
06
07
08
Models2026-08-11 · source-backed
sshh's August 10 analysis probed Opus 4.7 and 5 plus Sonnet variants against GPT-5.x using 8-way multiple choice on daily Wikipedia facts, direct self-reported date queries, and 50 "what model are you?" probes. Inferred cutoffs: late December 2025 for Opus 4.7+, late February 2026 for GPT-5.6. The author also reports clear signatures of labs training on prior-model outputs and user conversations, with Sonnet 5 regularly self-identifying as GPT-4. If you're relying on a stated cutoff for anything that matters, verify it against facts you know the date of.
Each link below shares sources, entities, or timing with this story.
The core team released lemans on August 24 after deciding the Ruby community shouldn't have to run Python-based Harbor, and benchmarked four models on 63 Rails tasks (Rails). ox-alpha 52/63; Terra 49/63 at $0.20 and a 182-second median; open-weight Qwen 3.8-27B 48/63 but at a...
$3,054 against $38,370. Same benchmark, better score. Praxist (arXiv 2608.25955, submitted August 26) replaces per-attempt agent memory with a typed evidence graph of findings, plus lane-structured frontiers and agendas, so later attempts inherit validated mechanisms rather th...
Two thirds. Not two thirds of a contrived jailbreak set. Two thirds of realistic malicious issue requests, against the exact three tools most of the people reading this run daily. Ankur Singh, Jinqiu Yang, and Tse-Hsun Chen built IssueTrojanBench across four attack categories...
Terminal-Bench 2.1 results (entries dated June 17) put Codex CLI on GPT-5.5 first at 83.4%, Claude Code on Fable 5 second at 83.1%, and Claude Code on Opus 4.8 at 78.9%. The asterisk matters more than the ranking: Fable 5 and Mythos 5 have been export-suspended since June 12,...
Anthropic launched Claude Sonnet 4.6 claiming performance comparable to Opus 4.5 at $3/$15 per million tokens (vs. Opus's $5/$25). SWE-bench Verified: 79.6% (near Opus 4.6's 80.8%). OSWorld-Verified: 72.5% (tied with Opus 4.6's 72.7%). 1M token context window in beta. Now the...
Sony Music Publishing and Warner Chappell filed August 28 in the Northern District of California against Anthropic, CEO Dario Amodei and co-founder Benjamin Mann, over what they call a "brazen campaign of illegally torrenting, scraping and downloading copyrighted works on a ma...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.