[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

![cortensor Avatar](https://lunarcrush.com/gi/w:24/cr:twitter::1802916496789929984.png) Cortensor [@cortensor](/creator/twitter/cortensor) on x 3865 followers
Created: 2025-07-23 02:56:19 UTC

🛠️ DevLog – CognitiveLevel Iteration: Capped Levels, SLA Categories, and Node-Aware Scheduling

As part of our ongoing CognitiveLevel iteration:

🔹 Level Cap Introduced:
We've capped levels to 0–5. Submitting recursive inference beyond Level X exceeds onchain tx size limits - so this is the new practical upper bound for now.

🔹 Session Integration Incoming:
With stable node categorization in place (Levels 0–5), we're integrating this into session creation UI/config. You'll soon be able to specify required SLA tiers when creating a session.

🔹 Toward SLA-Aware Scheduling:
We'e beginning with three SLA classes: low, medium, high.
Each session will select nodes closest to the requested SLA (i.e., based on CognitiveLevel) for better task quality and predictable performance.

This is a key step in bringing adaptive task routing and PoI enforcement deeper into the Cortensor execution pipeline.

#Cortensor #DevNet6 #CognitiveLevel #PoUW #PoI #DePIN #AIInfra #SessionConfig #SLA

![](https://pbs.twimg.com/media/Gwgp-iIbIAAlQ3u.jpg)

XXX engagements

![Engagements Line Chart](https://lunarcrush.com/gi/w:600/p:tweet::1947853251623457194/c:line.svg)

**Related Topics**
[rating agency](/topic/rating-agency)
[onchain](/topic/onchain)
[inference](/topic/inference)
[categories](/topic/categories)
[cortensor](/topic/cortensor)

[Post Link](https://x.com/cortensor/status/1947853251623457194)