Frontier labs spent the day arguing over who actually holds the lead, with fresh reviews of OpenAI’s newest model, a delayed flagship from Google, and an updated benchmark that keeps the crown in a familiar place.
GPT-5.6 Sol impresses on speed, stumbles on benchmarks
The first detailed reviews of OpenAI’s GPT-5.6 Sol landed, praising markedly faster coding and pricing roughly half that of Anthropic’s Fable 5, while flagging inconsistent results on some public reasoning benchmarks. The model remains gated to a small group of trusted partners under a government-reviewed rollout, so hands-on impressions still come mostly from early API access rather than the general public. Read the review.
Google pushes Gemini 3.5 Pro to July 17 for a full rebuild
Google DeepMind confirmed it is delaying Gemini 3.5 Pro to July 17, abandoning the previous Gemini 2.5 Pro architecture in favor of a ground-up redesign. The overhaul reportedly targets mathematical reasoning, SVG scene generation, and image quality as Google races to match GPT-5.6 and Anthropic’s Fable 5. More on the delay.
Latest intelligence index still favors Claude Opus 4.8
An updated snapshot of the Artificial Analysis Intelligence Index places Claude Opus 4.8 narrowly in front at 55.7%, with GPT-5.5 at 54.8% and Claude Opus 4.7 close behind. The tight 2.2-point gap at the top underlines how compressed the frontier has become, even as new releases arrive almost weekly. See the benchmark.
Meta restructures around AI, cutting thousands of roles
Meta began implementing a major AI-focused restructuring, reducing roughly 8,000 positions while reassigning thousands more staff to AI-centric teams. The move highlights how aggressively large platforms are reorganizing headcount and budgets around generative and agentic AI. Latest coverage.
What to watch: Google’s July 17 Gemini 3.5 Pro launch and any broadening of GPT-5.6 access will likely reset the leaderboard yet again.