BottleCap AI released ThinkingCap-Qwen3.8-27B as a fine-tune of Qwen3.8-27B to shorten reasoning traces. It cuts thinking tokens by 37.2% on average across 12 benchmarks while macro-average accuracy drops from 86.65% to 85.79% (0.86pp). It’s deployable as a drop-in on vLLM or SGLang with various quantized builds, but weights are gated and commercial use needs a BottleCap agreement beyond the small-business license.
Opus 5.5 was released on Tuesday as part of a wider update covering AI model releases, regulation proposals, and reported AI-involved incidents. At least 123 children were killed after outdated intelligence and an AI-assisted targeting workflow contributed to destroying an Iranian school. The write-up says this combination of faster AI processing and reduced human safety review should prompt tighter controls, while new lower-priced OpenAI model tiers (Sol and Luna) reflect shifting deployment priorities.
Liquid AI released an experimental DSpark draft model, LFM2.5-VL-DSpark, for its vision-language model LFM2.5-VL-3B to speed up speculative decoding without changing output quality.
Claude Opus 5.5 was released with updates to Claude Code, plus OpenAI’s GPT-6 Luna/Sol pricing changes also discussed. The Opus 5.5 release increased Claude Code’s five-hour limits by 20%. As a result, the default Claude Code “go-to agent” shifted to Claude Code for the author’s workflows while GPT-6 Sol’s lower cost showed mixed performance on long tasks.
Google DeepMind’s new chief said Gemini 4 is in refinement and aimed for an earlier release than the end of the year. The model is expected “much earlier” than that year-end target. As a result, Google plans to release an early post-training output first and keep iterating quickly based on the results.
Every AI story that matters,
in your inbox by 8am.
TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the
day in two minutes. Follow companies and topics for alerts, or get the
briefing in Slack. Free, no spam, unsubscribe anytime.
Reading TLDRocket needs no cookies, and the readership counts we rely on come from
our own cookieless analytics. Google Analytics is the exception: it sets cookies and
reports to Google, so it stays switched off until you allow it. You can change your
mind any time from “Cookie settings” in the footer.
Strictly necessary
Session security and form protection (tldrocket-session,
XSRF-TOKEN, 2 hours). The site cannot work without them,
so they need no consent.
Always on
Google Analytics 4 (_ga,
_ga_<id>, up to 2 years). Measures which
stories and sections readers use. Google acts as a third-party processor and may
store the data outside the EU. No advertising, no profiling, no data sold.
