In partnership with
News of the day
1. Anthropic just released Claude Opus 4.7. The model shows clear gains on complex coding, high-resolution vision and long-running agentic tasks, with a new xhigh effort level to better balance reasoning and latency. → Read more
2. Google launches Gemini app for macOS, offering seamless AI assistance directly from any screen without needing browser tabs. → Read more
3. Runway CEO proposes AI-driven shift: 50 films for $100M instead of one blockbuster, boosting output and hit potential. → Read more
4. DeepL launches voice-to-voice translation suite for meetings, conversations, and custom apps, plus an API for developers. → Read more
Our take
Hi Dotikers!
Yesterday, we were looking at Claude Code's Routines, those automations that run overnight while the developer sleeps. The bet was a risky one: delegating without supervision means hoping the engine lives up to the ambition. Anthropic just shipped the engine.
Claude Opus 4.7 has been available since this morning, and reading the announcement leaves a strange impression of industrial coherence. The model gains nearly 7 points over Opus 4.6 on SWE-bench Verified (87.6% versus 80.8%), outperforms GPT-5.4 and Gemini 3.1 Pro on most agentic benchmarks, and now accepts images three times more detailed than before. Pricing stays flat: 5 dollars per million input tokens, 25 per million output. A new xhigh effort level slots in between high and max, and the /ultrareview command opens a dedicated code review session, as if Anthropic decided its own model should proofread itself.
The detail that says more than the charts: several partners mention that the model now pushes its own arguments when it disagrees, checks its outputs, and flags when data is missing instead of inventing a plausible answer. Anthropic's messaging has quietly shifted. They are no longer selling a faster assistant, they are selling a collaborator that knows it can be wrong.
There is also what the announcement does not say out loud. Mythos Preview, the higher-tier model, scores 77.8% on SWE-bench Pro versus 64.3% for Opus 4.7. Anthropic is keeping it under wraps because of Project Glasswing and is testing its cyber safeguards on Opus 4.7 before releasing the rest. Reasonable translation: what we get today is already very good, and it is intentionally the second-best option in the catalog. The "state-of-the-art" label is deserved. It is also, incidentally, the price to pay for us to forget that the truly interesting version stays locked up.
Alex.
The IT strategy every team needs for 2026
2026 will redefine IT as a strategic driver of global growth. Automation, AI-driven support, unified platforms, and zero-trust security are becoming standard, especially for distributed teams. This toolkit helps IT and HR leaders assess readiness, define goals, and build a scalable, audit-ready IT strategy for the year ahead. Learn what’s changing and how to prepare.
Pic of the day
