Source notes
Source notes
Where these snapshots come from — Capability draws from Artificial Analysis, Usage Heat from OpenRouter routing, and Coding from SWE-bench-style score snapshots with explicit provenance.
RANK
Which model should you choose, and why?
Rank is the decision surface of SUPERCRZY. It separates intelligence, usage heat, and coding strength so one recommendation does not flatten three different kinds of judgment.
LIVE BOARD
Switch the lens, then read the board
This list is updated in real time. The data is sourced from the internet and is for reference only.
Primary source: Artificial Analysis
Snapshot: 2026 week 18
Use case: broad capability
Best quick take
Best quick take
Start with capability, then ask whether heat confirms it. If a model is hot but weak on coding, do not over-recommend it for engineering workflows.
What changes next
What changes next
Watch Chinese labs, agent-ready releases, and coding-specific jumps. The most meaningful movement usually shows up in heat and coding before it settles into broad capability rankings.
AGENT HANDOFF
Rank should end with a better decision, and sometimes that decision needs an agent.
Once a reader understands the model tradeoffs, the deeper step may be asking an agent to help compare workflows, narrow choices, or move into actual execution.