signal: week of 2026-07-08
2026-07-08
https://tinksoft.com/log/signal-2026-07-08/
--------------------------

A crowded week: two of the big three shipped major releases, and the third confirmed its timeline. Here's what actually landed.

- [Previewing GPT-5.6 Sol: a next-generation model](https://openai.com/index/previewing-gpt-5-6-sol/) — OpenAI's GPT-5.6 family (Sol, Terra, Luna) goes public July 9 after a government-coordinated limited preview, priced at $5/$30, $2.50/$15, and $1/$6 per million tokens respectively.

- [Anthropic restores Claude Fable 5 as US lifts export controls](https://www.tomshardware.com/tech-industry/artificial-intelligence/anthropic-restores-claude-fable-5-as-us-lifts-export-controls) — Fable 5 came back online July 1 after a 19-day export-control suspension, with a new safety classifier blocking the reported jailbreak in over 99% of cases and a shift to usage-based credits ($10/$50 per million tokens) as of July 7.

- [Claude Science, an AI workbench for scientists](https://www.anthropic.com/news/claude-science-ai-workbench) — Anthropic launched Claude Science in beta, a multi-agent workbench with 60+ preconfigured tools for genomics, proteomics, cheminformatics, and structural biology, plus a $30K credits grant program open through July 15.

- [Gemini 3.5 Pro cleared for July launch](https://www.techtimes.com/articles/319318/20260629/gemini-35-pro-cleared-july-launch-fable-5-nears-return-gpt-56-stays-locked.htm) — Google's Gemini 3.5 Pro is now confirmed for a July GA release after slipping twice from its original June target, featuring a 2M-token context window and a reasoning mode gated to the $250/month Ultra tier.

- [New Realtime models on the API: gpt-realtime-2.1 and gpt-realtime-2.1-mini](https://community.openai.com/t/new-realtime-models-on-the-api-gpt-realtime-2-1-and-gpt-realtime-2-1-mini/1385896) — OpenAI shipped gpt-realtime-2.1 and 2.1-mini on July 6 with 25% lower p95 latency, better alphanumeric recognition, and reasoning plus tool-use support in the mini model at no price increase.

- [GPT-5.6 prompt caching with explicit cache breakpoints](https://openai.com/index/previewing-gpt-5-6-sol/) — GPT-5.6 introduces explicit cache breakpoints and a guaranteed 30-minute minimum cache life, with cache writes billed at 1.25x the uncached input rate.