Under FSM metrics, GPT-3.5-Turbo outperforms GPT-4o-Mini, reversing that model's higher rank on coding-oriented leaderboards¶
Claim · All claims
Evidence · 1 study · designr2·q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly
Subclaims¶
Reading the evidence codes
q is the design tier, i the impact, n the sample, and the kind with its r
the rigour judged within that kind: what each code means.
This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.
q2 i? GPT-3.5-Turbo consistently outperforms GPT-4o-Mini under FSM metrics despite GPT-4o-Mini ranking above it (Z-score 0.30 vs −2.98) on the APXML coding leaderboard. → Xiaozao Wang 2026
Evidence¶
Xiaozao Wang 2026¶
Xiaozao Wang, Zhewei Wang, Hongyi Wen. (2026). Evaluating Interactivity: Toward Automated Assessment of AI-Generated Explorable Explanations. https://arxiv.org/abs/2606.31012 doi:10.48550/arxiv.2606.31012
q2 · i? · design · r2
Results-section comparison across the six generated-model corpus. The authors attribute the reversal to GPT-4o-Mini producing functionally correct interfaces that under-specify interaction states, while GPT-3.5-Turbo more consistently externalizes interaction structure, yielding higher FSM coverage and behavioral coherence.
"Under FSM metrics, however, GPT-3.5-Turbo consistently outperforms GPT-4o-Mini despite its lower nominal capacity."
Discussion¶
Related Claims¶
- FSM-based evaluation discriminates interactive quality across models, with GPT-5 Mini scoring highest and most tiers significantly separated — related
- Model selection experiments show newer models are not strict improvements, with component-specific effects on quality metrics — a broader claim this one bears on
- GPT-4o-mini natural language feedback was overwhelmingly helpful and never harmful in quality annotation — related
- The OpenAI gpt-4o reference model outperformed all quantised models on human-rated coherence, fluency, relevance and cultural appropriateness, serving as an upper benchmark — related
- FSM-based evaluation aligns with human judgments of interactivity, functional correctness, and visual quality more strongly than VLM and unit-test baselines on balanced data — related
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
fsm-gpt35-outperforms-gpt4o-mini-reversal |
title |
Under FSM metrics, GPT-3.5-Turbo outperforms GPT-4o-Mini, reversing that model's higher rank on coding-oriented leaderboards |
description |
Under FSM metrics, GPT-3.5-Turbo outperforms GPT-4o-Mini, reversing that model's higher rank on coding-oriented leaderboards |
status |
review |
evidence_strength |
mixed |
newest_evidence |
2026 |
recheck_by |
2029 |
generated |
by process:wiki-ingest on 2026-10-10 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — xiaozao-wang-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.