Cognition adds SpaceXAI as a fourth frontier model in Devin Pro, three days after SWE-2 default swap

WHY IT MATTERS

Devin Pro's frontier model lineup is becoming an active lever Cognition keeps adjusting rather than a settled feature list. Watch whether Cognition keeps adding model providers at this tighter pace or reverts to its roughly three-week rhythm.

BY THE NUMBERS

• 3 days after Cognition's last change (SWE-2 swap, Sep 14)
• 4 real changes since tracking began Jul 6, 2026
• ~22-day average gap between changes, historically
• Feature changes make up 15.7% of all tracked pricing moves (867/5515)

THEIR RATIONALE

No explicit rationale is stated, but the timing tracks Cognition's recent pace of coding-model iteration: this follows just three days after Cognition swapped its platform-wide default model to SWE-2, part of a run of four feature changes to the page since tracking began in July.

THE SIGNAL

This is Cognition's fourth pricing-page change since being tracked, all feature changes events, at an average cadence of about 22 days — but the SpaceXAI addition landed just 3 days after the SWE-2 default swap, compressing that cadence into a burst. Feature-lineup changes are the single most common move across the index (867 of 5515 changes tracked in six months, 15.7%), so expanding model choice on Pro is squarely inside category norms, even if the 3-day turnaround is fast for Cognition itself.

IN CONTEXT

Cognition's pricing page has moved fast since Pulse Index started tracking it on Jul 6, 2026 — four changes in barely two months, all clustered around its coding-model lineup. This one lands just three days after Cognition swapped its default model to SWE-2, and follows two earlier model upgrades over the summer, July's Pro-plan SWE 1.7 bump and the Compare Plans update the week before. Feature-lineup tweaks like this are the most common move in the index — 15.7% of all changes tracked in the last six months — with frequent movers like HubSpot and Datadog doing the same. Cognition's version of that pattern: rapid churn in which AI models it bundles into Devin Pro, on a tightening timeline.