Evidaxis · Pilot 2026
An observatory of momentum in open AI systems.
We measure open-source and research-native AI systems by the rate of change of their public signals — not their size. A system is recognized as rising only when two independent axes move together. Every measurement is released to the public domain under CC0.
The instrument
Two axes. One gate.
Each system is measured on independent axes that move on different timescales. Development velocity is the log-slope of commit activity — a fast, leading signal. Citation momentum is the log-slope of yearly citations to the system’s paper — a slow, confirming signal. Each is normalized within its cohort, so “rising” means rising relative to peers, not simply large.
A system earns the Rising badge only when both axes are rising at once. Stars and raw size are never scored. The gate publishes only positive recognition — there is no “worst” list.
Rising = development velocity ∧ citation momentum, both above the cohort median.
State of snapshot 90e607b982fa
Convergence is rare — and that is the point.
At a baseline snapshot, very few systems show both axes rising at once: a fast-moving codebase is often years ahead of its citations, and a genuinely new system has no citation history yet. That scarcity is by design — the signal sharpens as the weekly time-series accumulates. Below, the systems whose one axis is already rising: candidates on the way to convergence.
Both axes rising. The full badge.
One axis rising, awaiting a second.
Measured for cohort context.
Watch — candidate signals
One independent axis is rising. Not yet a convergence badge.
| # | System | Status | Momentum | Dev velocity | Citation momentum |
|---|---|---|---|---|---|
| 1 | Boltz e_9VD4D2XPK7H | Watch | 80.0 | ||
| 2 | Cline e_P4YGBAFMV94 | Watch | 60.5 | no paper | |
| 3 | dora-rs e_Y31RBTP1VWD | Watch | 60.4 | no paper | |
| 4 | Graphormer e_8HVZ2FE02WG | Watch | 60.4 | ||
| 5 | Isaac Lab e_N0ZEZQYXT6K | Watch | 41.6 |