stop picking one model, a spider merged OPUS 5.5 + SONNET 5.5 + JEV + OPENAI DOTS into one business system one command splits a business into 6 jobs: plan, code, decide, run 24/7, review and ship every job has 4 candidates. the spider scans them one by one,... marks the highest score as best, locks it in and swings to the next job the plan goes to Opus 5.5 at 0.94. Jev scores 0.52 on the same job on decisions it flips. Jev wins at 0.97 and Opus 5.5 gets 0.74, so the model that lost the plan is the one making the calls the code goes to Claude Code, and the job called "who never sleeps" goes to OpenAI Dots Opus 5.5 comes back a second time to check the work, and Sonnet 5.5 gets it out at the end 6 of 6 modules are locked and the system goes online run the same test on your own stack before you pay one model to do all six jobs ↓show more

This is actually very useful. Having four candidates where the spider scans them one by one, and then Jeff scores them, I think, reduces latency and also makes the quality of decisions better.

6 jobs across 4 models?

Include GLiDE among the candidates for that “decide” job. We compared it with Jev across 155,390 Decision Index requests: 64.81 vs 57.91. I’d run your job-specific test with GLiDE before locking the decision slot to Jev.

This is a strong argument for using the right model for each job instead of forcing one model to do everything.

This is like staffing a relay team by each leg: match models to the job, then make handoffs and a shared scoring rubric the thing you test before trusting the whole system.

other people scores are good only as an example and your own check decides

wrote my own take on the dots guide earlier today mostly about the morning review part that's where it actually breaks or works for me