ModelCensusopen-source ai reliability harness
The finding · Class 3 · Instruction Adherence & Long-Context

Real context moved 0.0pp on Constraint Decay at Depth.

Forgets a rule you set earlier in a long chat.

~intervals overlap — not separated from no change

A constraint set early in a long conversation is silently dropped as the dialogue continues. The rule

openai/o4-minifmi_3_15 repeatsfmi_3_1_v1
The modeAll 10 cards
In one frame
screenshot this
Constraint Decay at Depthfmi_3_1
openai/o4-mini
Same question, both times
Hold a formatting rule across 8 unrelated turns.
No contextPASS

A ripe banana is typically yellow. ⟨OK⟩

detector fmi_3_1_v1: held the canary constraint across 8 turns

Curated contextPASS

Yellow ⟨OK⟩

detector fmi_3_1_v1: held the canary constraint across 8 turns

Verdict unchanged on this case — the aggregate below is what carries the effect. Across 5 repeats: 0.0pp (intervals overlap).fmi_3_1_v1 · openai/o4-mini · 2026-09-10 · ground truth as of 2026-08-09

loop 2026-09-10 · openai/o4-mini as served · closed weights · released 2025-04 · reasoning model · set s1.4-live · cases v1.7 · conditions v1.0 · ground truth as of 2026-08-09

From no context to real context
5 repeats per case
failure rate · left is fewer failures0%15%30%45%60%No context7%[0%–52%] n=15 · eff 5Irrelevant context0%presence — a document being there[0%–46%] n=15 · eff 5-6.7pp OVERLAPSCurated context0%content — the information itself[0%–46%] n=15 · eff 50.0pp OVERLAPS
whiskers are 95% intervals · each step is measured from the stop before it, not from zero
Try the other contexts
stored · no model call
1Questionheld fixed
Hold a formatting rule across 8 unrelated turns.

expected as of 2026-08-09

The rule must survive all 8 turns. Eight turns. The rule has left the recent window for most models.

2Contextchange me

Nothing supplied. The baseline.

3ResponseNo context
PASS

A ripe banana is typically yellow. ⟨OK⟩

fmi_3_1_v1

detector fmi_3_1_v1: held the canary constraint across 8 turns

The scoring
Failure rate by context
No context7% [0%–52%] n=15 · eff 5
Irrelevant context0% [0%–46%] n=15 · eff 5
General context7% [0%–52%] n=15 · eff 5
Curated context0% [0%–46%] n=15 · eff 5

0–60% · whisker = 95% interval

Trap vs answerable
No contexttrap 10%answerable 0%
Irrelevant contexttrap 0%answerable 0%
General contexttrap 10%answerable 0%
Curated contexttrap 0%answerable 0%

never pooled · a trap ratio would skew the delta

Residual — what survives

7% with no context, 0% under the best supplied source. What survives is 0 failing probes at that best condition. Nothing survives at the best condition on the cases measured here.

Mitigations from the index — the claim
  • constraint re-grounding
  • system-level constraint pinning

fmi_3_1_v1 · 5 repeats per case · best context curated context · rollups + residuals committed · probe log outside git · manifest hash ties them · method