Nine New Models in Three Weeks, and You Were Already Using Most of Them

Claude Sonnet 5 replaced Sonnet 4 in the Operators seat about a week and a half ago.

We never announced it. I only worked that out while writing this post.

That is not great changelog hygiene on my part. But it is a decent illustration of how model updates are supposed to work here, so I am leaving the embarrassment in.

Between the start of July and yesterday, nine new models landed in Suprmind. Three from OpenAI, three from Anthropic, two from Google, one from x.AI. Where a new release replaced an older default, it took the seat automatically. No migration notice, no action required, no settings to reconcile. You kept working and the models under you got newer.

What landed

Google shipped its newest generation on July 21 and it was running on Suprmind days later. Grok 4.5 arrived the same day as a selectable option, though Grok 4.3 stays the default across all three teams for now.

Gemini 3.1 Pro keeps the A-team Gemini seat. Sonar Reasoning Pro holds all three Perplexity seats, unchanged. Claude Haiku 4.5 still runs Daily Drivers.

What I am not going to claim

That you will feel all nine of these on every prompt.

For a large share of everyday work, the gap between Sonnet 4 and Sonnet 5 is not something you would spot without looking for it. Same with Gemini 3.5 Flash to 3.6 Flash. What moves is the ceiling, not the floor, and the ceiling only matters on the work that pushes against it.

We also have not run our own benchmarks on this batch yet. What I can tell you is what the providers published and what we have seen in production over the past few weeks. When we have our own numbers, they go in the Multi-Model Divergence Index, not in a changelog post.

Where the models actually live

Suprmind runs three AI Teams, and each team has one seat per provider.

The A-team is deep deliberation for strategic decisions, novel problems, and high-stakes calls. Operators is the workhorse for real tasks that do not need maximum brainpower at maximum cost. Daily Drivers covers speed and large context for parsing, structured extraction, and data work.

Every one of those seats is a dropdown, and every dropdown shows that provider’s full lineup. You can put Claude Opus 5 in Daily Drivers if you want the heavy model doing your parsing. You can drop Claude Haiku 4.5 into the A-team if you are running high volume and care more about runway than depth. Nothing stops you. Defaults hold until you override them.

If a roster stops making sense, each team has its own Reset team button. It resets that team only. Your other two stay exactly as you left them, so you can experiment on Daily Drivers without touching the A-team you spent time tuning.

All of it lives in Account > Settings > Usage Control.

Anything else you need to know

No. Pricing did not change. Usage capacity did not change. If you had manually picked a model that got retired, your pick carried forward to its successor on its own.

The rosters are worth ten minutes if you have opinions about which model should be doing what. If you do not have opinions about that, the defaults are the ones we would pick for you anyway.

Open Usage Control

Radomir Basta CEO & Founder

Radomir Basta builds tools that turn messy thinking into clear decisions. He is the co founder and CEO of Four Dots, and he created Suprmind.ai, a multi AI decision validation platform where disagreement is the feature. Suprmind runs multiple frontier models in the same thread, keeps a shared Context Fabric, and fuses competing answers into a usable synthesis. He also builds SEO and marketing SaaS products including Base.me, Reportz.io, Dibz.me, and TheTrustmaker.com. Radomir lectures SEO in Belgrade, speaks at industry events, and writes about building products that actually ship.

See Full Bio