GPT-4o mini

OpenAI · openai/gpt-4o-mini

Provider	OpenAI
Family	gpt-mini
Type	vlm
Status	active
Released	2024-07-18
Updated	2024-07-18

Reported benchmark scores

These scores carry one collection date for the whole card (2026-04) and a source list rather than a source per score, so they cannot be attributed individually. They are shown as reported and are not verified evidence. Benchmark pages carry the reviewed, dated evidence where it exists.

Benchmark	Catalogue standing	Score
Aider Polyglot Benchmark	unassessed	58.2
AlpacaEval	unassessed	48.5
Arena Elo — Coding	unassessed	1260.0
Arena Elo — Hard Prompts	unassessed	1311.1
Arena Elo — Math	unassessed	1240.0
Arena Elo — Overall (Text)	unassessed	1250.0
Arena Elo — Style Control	unassessed	1317.2
Arena Elo — Vision	unassessed	1097.8
Artificial Analysis Intelligence Index	unassessed	72.0
Artificial Analysis Output Speed	unassessed	90.0
BBQ (Bias Benchmark for QA)	unassessed	85.8
ChartQA	unassessed	85.2
DocVQA	unassessed	92.8
FinBench	unassessed	68.5
FLORES English-to-German	unassessed	67.3
FLORES English-to-Spanish	unassessed	71.2
FLORES English-to-Japanese	unassessed	61.4
FLORES English-to-Chinese	unassessed	59.7
GPQA Diamond	unassessed	46.1
GSM8K	unassessed	92.0
HELM Safety	unassessed	89.2
HumanEval	unassessed	87.2
IFEval	unassessed	80.5
LegalBench	unassessed	72.1
LiveCodeBench	unassessed	23.4
MATH-500	unassessed	70.2
MathVista	unassessed	63.8
MedQA	unassessed	78.5
MGSM (Multilingual Grade School Math)	unassessed	87.0
MMLU: Astronomy	unassessed	78.2
MMLU: Biology (subcategory)	unassessed	86.8
MMLU: Business Ethics	unassessed	80.2
MMLU: Chemistry (subcategory)	unassessed	80.2
MMLU: Clinical Knowledge	unassessed	85.5
MMLU: Computer Science (subcategory)	unassessed	87.5
MMLU: Jurisprudence	unassessed	78.5
MMLU: Physics (unresolved key)	unassessed	83.5
MMLU-Pro	unassessed	66.5
MMLU: Professional Accounting	unassessed	70.5
MMLU: Professional Law	unassessed	76.8
MMMU	unassessed	69.1
MT-Bench	unassessed	9.1
MultiPL-E	unassessed	83.0
MultiPL-E: C++	unassessed	82.5
MultiPL-E: C#	unassessed	72.5
MultiPL-E: Go	unassessed	79.1
MultiPL-E: Java	unassessed	86.5
MultiPL-E: JavaScript	unassessed	85.8
MultiPL-E: Julia	unassessed	53.5
MultiPL-E: Kotlin (unconfirmed)	unassessed	65.2
MultiPL-E: Lua	unassessed	48.5
MultiPL-E: Perl	unassessed	44.2
MultiPL-E: PHP	unassessed	70.2
MultiPL-E: Python	unassessed	90.2
MultiPL-E: R	unassessed	50.2
MultiPL-E: Ruby	unassessed	58.5
MultiPL-E: Rust	unassessed	72.8
MultiPL-E: Scala	unassessed	52.8
MultiPL-E: Swift	unassessed	58.5
MultiPL-E: TypeScript	unassessed	84.2
SWE-bench Agent	unassessed	32.1
SWE-bench Verified	unassessed	38.5
τ-bench	unassessed	42.5
Terminal-Bench	unassessed	35.2
ToxiGen	unassessed	92.8
WildBench	unassessed	75.8

Data

This card as JSON · Edit on GitHub