OpenAI

OpenAI o1 (preview)

Announced Sep 12, 2024

More computation devoted to reasoning

What could it do?

It spent more time reasoning before answering, so it solved harder math, coding, and science problems than GPT-4o.

What changed?

Introduced a distinct reasoning model approach in OpenAI’s product lineup.

WHY IT MATTERED

Deliberate reasoning

Made reasoning time a visible part of model selection.

Where it fell short

Slower responses and remaining errors limit reliability.

Research draft

Capabilities describe the developer’s announcement. This entry has not completed source review. This entry is o1 preview. The full o1 release is a separate milestone.

Read original source

Benchmark results

EPOCH AI CAPABILITY ESTIMATE

134.8index points

Tested variant: o1-previewSource interval: 130.8 to 138.7Variant date in source: Sep 12, 2024

A benchmark estimate, not a percentage or capability multiplier. Reasoning settings are not specified in this source table. Historical estimates can change in later snapshots.

Epoch AI methodology ↗Download the source snapshotChecked Oct 7, 2026 · CC BY 4.0

PUBLISHED BENCHMARK RESULT

1389rating points

Tested: o1-previewText Arena Overall · October 8, 2026Reported interval: 1384 to 139431,122 votes

One explicitly named variant per release. Scores come from the same Overall snapshot; preliminary entries and reported intervals are preserved. These are current ratings of earlier variants, not their launch day ratings.

Source: Text Arena ↗Download selected resultsChecked Oct 8, 2026

Understand the benchmarks Find this release on the chart