o1-mini

OpenAI

Verified against Artificial Analysis · Jul 13, 2026

Capability

percentile index

Price /M

$0 / $0

in / out

Context

—

Avg score

54.0

6 benchmarks

AA Intelligence

14.0

index

Weights

Closed

Released

Sep 12, 2024

About

o1-mini is a proprietary model in the o1 family from OpenAI. Benchmarked on 6 evals, averaging 54.0.

Benchmark scores · 6

avg 54.0 — higher is better

MATHartificial-analysis · 2026-07

94.4

MMLU-Proartificial-analysis · 2026-07

74.2

GPQA Diamondartificial-analysis · 2026-07

60.3

LiveCodeBenchartificial-analysis · 2026-07

57.6

SciCodeartificial-analysis · 2026-07

32.3

Humanity's Last Examartificial-analysis · 2026-07

4.9

The research behind this model

all papers mentioning it →

MiMo: Unlocking the Reasoning Potential of Language Model -- From Pretraining to Posttraining1 cites · May 12, 2025 Mathematical Reasoning via Intervention-Based Time-Series Causal Discovery Using LLMs as Concept Mastery SimulatorsMay 8, 2026

Mentions matched by name in title/abstract — from the arXiv + HF daily corpus.

Details

PublisherOpenAI

Familyo1

Parameters—

License—

WeightsClosed

ReleasedSep 12, 2024

History

full ledger →

Capability over time

Append-only ledger — every observed change to this model's numbers.

API providers

Artificial Analysis$0/M