The Model Index
Up to four models

GPT-5.4 Mini vs Kimi K2.5

On the index (GPT-4 = 100), GPT-5.4 Mini scores 147 and Kimi K2.5 145. 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 GPT-5.4 Mini
Presets
GPT-5.4 Mini
147131–163
WeightsClosed
ReleasedMar 2026
Price in / out$0.75 / $4.5
Context400K
Core tests taken5 of 9
Kimi K2.5
145130–161
WeightsOpen
ReleasedJan 2026
Price in / out$0.50 / $2.5
Context262K
Core tests taken5 of 9

GPT-5.4 Mini and Kimi K2.5 are too close to call (2 points apart, within the uncertainty), ahead on 2 of 3 shared core tests, at 1.7× the price.

Test by testthird-party results

GPT-5.4 MiniKimi K2.5
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Epoch AI)GPQA Diamond (Epoch AI)GPQA Diamond (Vals AI)GPQA Diamond (Vals AI)MMLU ProMMLU ProMMMU ProMMMU ProSimpleQA VerifiedSimpleQA VerifiedMATHSAIME-style maths (OTIS mock)AIME-style maths (OTIS mock)FrontierMath (2025 set)FrontierMath (2025 set)FrontierMath tier 4 (2025 set)FrontierMath tier 4 (2025 set)CODINGCode MigrationCode MigrationLiveCodeBenchLiveCodeBenchSWE-benchSWE-benchVibe Code Bench v1.1Vibe Code Bench v1.1WeirdMLWeirdMLPROFESSIONAL WORKFinance Agent (v2)Finance Agent (v2)CorpFin v2CorpFin v2EMBEMBLegal Research BenchLegal Research BenchMortgageTaxMortgageTaxTaxEval v2TaxEval v2NOVEL PROBLEMSARC-AGI-2ARC-AGI-2ARC-AGI-1ARC-AGI-1
Every result as a table
TestGPT-5.4 MiniKimi K2.5
Reasoning & knowledge
GPQA Diamond (Epoch AI)86.9%87.6%
GPQA Diamond (Vals AI)83.1%84.1%
MMLU Pro84.5%85.9%
MMMU Pro79.3%84.3%
SimpleQA Verified29.4%34.3%
Maths
AIME-style maths (OTIS mock)88.9%92.2%
FrontierMath (2025 set)28.3%27.9%
FrontierMath tier 4 (2025 set)2.1%4.2%
Coding
Code Migration12.9%7.0%
LiveCodeBench81.5%83.9%
SWE-bench73.0%70.0%
Vibe Code Bench v1.148.0%17.5%
WeirdML60.3%45.6%
Professional work
Finance Agent (v2)45.4%35.8%
CorpFin v260.9%68.3%
EMB45.4%28.5%
Legal Research Bench12.5%15.9%
MortgageTax63.5%66.5%
TaxEval v271.2%74.2%
Novel problems
ARC-AGI-218.9%11.8%
ARC-AGI-163.7%65.3%

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

This page as a chartDownload PNG