Evidaxis

FlashInfer

repo
Inference & Serving · n = 14 · flashinfer-ai/flashinfer
2 of 3 axes measurable · conf medium
convergence gate
Convergence gate: Gate: computed weekly across the cohort; no convergence event this week Development velocity: rising vote Citation momentum: not measurable ∅ Direct-dependents momentum (deps.dev): measured ×
gate
1 of 3 axes

Gate: computed weekly across the cohort; no convergence event this week

Gate: computed weekly across the cohort; no convergence event this week
position in cohort distribution one of 14 in cohort
percentile 77 within cohort p25 p50 p75 p0 p100
percentile
p77
momentum averages present-axis z-scores; two- and three-axis scores are not strictly comparable
velocity z
velocity z gaugevelocity z+2.18 within cohort, log-slope +0.061/wk.0+2.18
velocity z
+2.18 σ
slope
+0.061/wk
log-slope +0.061/wk
citation z Citation axis: insufficient source coverage (source returned too few observations).
direct-dependents z
direct-dependents z gaugedirect-dependents z-0.97 within cohort, log-slope +0.022/point.0-0.97
direct-dependents z
-0.97 σ
slope
+0.022/point
Direct-dependents axis: measured.
cohort
14
systems measured
weekly commits
89.3
Development velocity axis: measured.
total citations
n/a
Citation axis: insufficient source coverage (source returned too few observations).
momentum
57.6
percentile p77
commit heatmap, last 26 weeks89.3/wk
weekly commits, 26 weeks26 weeks of commits, peak 240 in a week, 26 active weeks, 1514 total.week 1: 33 commitsweek 2: 13 commitsweek 3: 38 commitsweek 4: 33 commitsweek 5: 28 commitsweek 6: 20 commitsweek 7: 27 commitsweek 8: 32 commitsweek 9: 12 commitsweek 10: 33 commitsweek 11: 31 commitsweek 12: 50 commitsweek 13: 44 commitsweek 14: 48 commitsweek 15: 48 commitsweek 16: 53 commitsweek 17: 65 commitsweek 18: 53 commitsweek 19: 45 commitsweek 20: 70 commitsweek 21: 100 commitsweek 22: 91 commitsweek 23: 131 commitsweek 24: 240 commitsweek 25: 163 commitsweek 26: 13 commits
weeks
26
peak/wk
240
active
26
total
1514
citations by yearCitation axis: insufficient source coverage (source returned too few observations).

FlashInfer holds an Evidaxis momentum score of 57.6 of 100 on methodology m4, snapshot 2026-10-10, measured within the Inference & Serving cohort (n = 14). Axis coverage: 2 of 3 axes measurable. Gate: computed weekly across the cohort; no convergence event this week

Derived signals

second-order reads, computed from the published series

Three signals ship when the commit series is statistically sufficient; the rest are held in a reserved state until the data supports them. A reserved signal shows its reason, never a fabricated number. How these are computed.

recency-weighted swing trailing 13-week vs prior 39-week log-slope ratio +2.68, regime accelerating. prior slope +0.017, recent slope +0.046 per week. +0.017prior 39w+0.046recent 13w+2.68×
swing
+2.68×
regime
accelerating
windows
13w vs 39w
snapshot 1 of 1 · · m4
commit-axis Sharpe log-linear slope over residual std, value +0.065 with 95% CI +0.044 to +0.087, 50 degrees of freedom. CI excludes zero.0+0.06595% CI +0.044, +0.087+0.044+0.087
sharpe
+0.065
95% ci
+0.044, +0.087
dof
50
snapshot 1 of 1 · · m4
commit changepoint52 weekly commit points. Changepoint at week 50 (2026-09-26), slope change -114.715 commits per week.2026-09-26
date
week
50 of 52
Δslope
-114.715/wk
snapshot 1 of 1 · · m4
code-leads-citation lag reserved

reserved: citation series too short / axis absent for cross-correlation

status
reserved
value
insufficient data
snapshot 1 of 1 · · m4
axis alpha vs cohort reserved

reserved: cohort N too small for regression (power ~0)

status
reserved
value
insufficient data
snapshot 1 of 1 · · m4
gate ETA reserved

reserved: needs snapshot history (1 snapshot exists)

status
reserved
value
insufficient history
snapshot 1 of 1 · · m4
methodology gate snapshot 1 of 1 · · m4
commit weeks
52
citation years
2
shipped
3
reserved
3

shipped

  • recency-swing
  • commit-sharpe
  • commit-changepoint

reserved

  • code-leads-citation-lag
  • axis-alpha
  • gate-eta
snapshot 1 of 1 · · m4

Reconstructed history

FlashInfer reconstructed weekly commits51 points, range 8 to 100 commits/wk.
min
8 commits/wk
max
100 commits/wk
latest
91 commits/wk
points
51

Weekly commits over the past year, reconstructed from git history (GitHub commit_activity). This is reconstructed, not a point-in-time capture. Evidaxis captures the Type-2 signals (watchers, issues, dependents) live and never backfills them; those are the moat.

Measurement states

FlashInfer is under weekly measurement. History: 12 weekly observations; minimum 4. Development velocity axis: measured. Citation axis: insufficient source coverage (source returned too few observations). Direct-dependents axis: measured. Gate: computed weekly across the cohort; no convergence event this week

Sources

flashinfer-ai/flashinfer development-velocity axis
OpenAlex paper citation-momentum axis

Others in Inference & Serving

AIBrix Inference & Serving
1 of 3 axes measurable
66
p100
vel +1.30 cit no axis deps no axis
axis dials
+1.30 n/a n/a
KTransformers Inference & Serving
1 of 3 axes measurable
42
p38
vel -0.62 cit no axis deps no axis
axis dials
-0.62 n/a n/a
llama.cpp Inference & Serving
2 of 3 axes measurable
53
p69
vel -0.16 cit no axis deps +0.67
axis dials
-0.16 n/a +0.67
llm-d Inference & Serving
1 of 3 axes measurable
33
p8
vel -1.38 cit no axis deps no axis
axis dials
-1.38 n/a n/a
LMCache Inference & Serving
1 of 3 axes measurable
44
p46
vel -0.48 cit no axis deps no axis
axis dials
-0.48 n/a n/a
LMDeploy Inference & Serving
1 of 3 axes measurable
27
p0
vel -1.82 cit no axis deps no axis
axis dials
-1.82 n/a n/a

Full Inference & Serving cohort →

Direct-dependents momentum (deps.dev)

Direct-dependents axis: measured.

Coverage state
scored
Cohort z
-0.97
OLS slope
+0.022
Theil-Sen slope
+0.022
Latest direct dependents
15
Confirmed-clean points
26
Reconstructed history points
15
As-of partition
2026-09-28
Fragility veto
inactive
Rising vote
no

This axis measures the trajectory of a system's own packages: declared in its own tree, published under that name, and not attributed by deps.dev to another repository. A package with no repository link is unverified and stays. The self-name exclusion list is fixed at the v2h.1 panel. Systems with no admitted package are not measured on this axis. Points that cannot be derived exactly are withheld, and the series restarts after a withheld point.

The earliest weekly points of each series were read from deps.dev's BigQuery history on 2026-07-21; later points are weekly captures. m4 does not replay those partitions: each point is the stored system total, one package's unique_direct, or a complete sketch union, and a point that is not exact is withheld. Each record reports how many of its points are reconstructed history.

Source: deps.dev (Google), CC-BY 4.0. Modified: aggregated to system-level counts and momentum scores.

Axis definition · Snapshot axis correlations

entity_id e_VMQ8QRRDWVS methodology m4 snapshot ad5e404d4f82 period 2026-w41 record JSON ↓ source snapshot

Score derived from public data on Evidaxis methodology m4, never edited by hand. Cite as: “Evidaxis, FlashInfer, entity e_VMQ8QRRDWVS, snapshot 2026-w41.”

Durable reference (format-independent): urn:evidaxis:claim:e_VMQ8QRRDWVS:m4:2026-10-10