The Model Index
Up to four models

GLM-5.1 vs GPT-5.4 Mini

On the index (GPT-4 = 100), GLM-5.1 scores 151 and GPT-5.4 Mini 147. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.

Next to GLM-5.1
Presets
GLM-5.1
151134–167
WeightsOpen
ReleasedApr 2026
Price in / out$1.4 / $4.4
Context205K
Core tests taken5 of 9
GPT-5.4 Mini
147131–163
WeightsClosed
ReleasedMar 2026
Price in / out$0.75 / $4.5
Context400K
Core tests taken5 of 9

GLM-5.1 and GPT-5.4 Mini are too close to call (4 points apart, within the uncertainty), ahead on 1 of 3 shared core tests, at a similar price.

Test by testthird-party results

GLM-5.1GPT-5.4 Mini
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Epoch AI)GPQA Diamond (Epoch AI)GPQA Diamond (Vals AI)GPQA Diamond (Vals AI)MMLU ProMMLU ProSimpleQA VerifiedSimpleQA VerifiedMATHSFrontierMath (tiers 1-3)FrontierMath (tiers 1-3)AIME-style maths (OTIS mock)AIME-style maths (OTIS mock)Chess puzzlesChess puzzlesFrontierMath (2025 set)FrontierMath (2025 set)FrontierMath tier 4 (2025 set)FrontierMath tier 4 (2025 set)CODINGCode MigrationCode MigrationLiveCodeBenchLiveCodeBenchProgramBenchProgramBenchSWE-benchSWE-benchVibe Code Bench v1.1Vibe Code Bench v1.1WeirdMLWeirdMLPROFESSIONAL WORKFinance Agent (v2)Finance Agent (v2)CorpFin v2CorpFin v2Legal Research BenchLegal Research BenchTaxEval v2TaxEval v2
Every result as a table
TestGLM-5.1GPT-5.4 Mini
Reasoning & knowledge
GPQA Diamond (Epoch AI)89.9%86.9%
GPQA Diamond (Vals AI)84.5%83.1%
MMLU Pro86.9%84.5%
SimpleQA Verified34.0%29.4%
Maths
FrontierMath (tiers 1-3)36.8%51.2%
AIME-style maths (OTIS mock)93.3%88.9%
Chess puzzles20.0%23.0%
FrontierMath (2025 set)33.5%28.3%
FrontierMath tier 4 (2025 set)12.5%2.1%
Coding
Code Migration25.8%12.9%
LiveCodeBench81.4%81.5%
ProgramBench0.0%0.0%
SWE-bench76.4%73.0%
Vibe Code Bench v1.131.5%48.0%
WeirdML57.1%60.3%
Professional work
Finance Agent (v2)44.8%45.4%
CorpFin v264.5%60.9%
Legal Research Bench27.9%12.5%
TaxEval v271.2%71.2%

Tests at least two of these models have taken. Bold rows are the core tests.

This page as a chartDownload PNG