GPT-5 Mini

ModelActive

by OpenAI · family “gpt-mini”

GPT-5 Mini is a compact version of GPT-5, designed to handle lighter-weight reasoning tasks. It provides the same instruction-following and safety-tuning benefits as GPT-5, but with reduced latency and cost.... (description from the OpenRouter listing)

in: textin: imagein: fileout: textReasoningTool useStructured output

Epoch Capabilities Index

145.5 #78 of 268

90% CI 143.5 – 147.0

Listing ↗

Released

7 Aug 2025

Context window

400K tokens

Max output

128K tokens

Input $ / 1M tokens

$0.25

Output $ / 1M tokens

$2.00

Cached input $ / 1M

$0.025

Each benchmark shown separately with its own source

Benchmark results (17)

All benchmarks

Benchmark	Domain	Score	vs best recorded	Setting	Run	Source
LMCA	agents	34.2%	52%	high	—	External ↗
DTBench	reasoning	80.5%	82%	high	—	External ↗
Mystery Game Puzzles	games	5.0% ±2.2	6%	high	27 Aug 2026	Epoch ↗
FrontierMath-Tiers-1-3-v2-Private	math	46.7% ±3.0	50%	high	12 Jun 2026	Eval log ↗
FrontierMath-Tier-4-v2-Private	math	12.2% ±5.2	12%	high	12 Jun 2026	Eval log ↗
ProofBench	math	9.0%	9%	high	—	External ↗
Chess Puzzles	games	30.0% ±4.6	42%	high	7 Aug 2026	Eval log ↗
SimpleQA Verified	knowledge	21.6% ±1.3	29%	high	10 Aug 2026	Epoch ↗
FrontierMath-Tier-4-2025-07-01-Privatesuperseded	math	6.3% ±3.5	13%	high	30 Oct 2025	Epoch ↗
ARC-AGI-2	reasoning	4.4%	5%	high	—	External ↗
FrontierMath-2025-02-28-Privatesuperseded	math	27.2% ±2.6	52%	high	13 Nov 2025	Epoch ↗
VPCT	multimodal	39.0%	43%	high	—	External ↗
WeirdML	coding	52.7% ±0.0	56%	high	—	External ↗
OTIS Mock AIME 2024-2025	math	86.7% ±4.3	87%	high	30 Oct 2025	Eval log ↗
GPQA diamond	science	75.0% ±2.3	78%	high	30 Oct 2025	Eval log ↗
MATH level 5	math	97.8% ±0.3	100%	high	30 Oct 2025	Epoch ↗
ARC-AGI	reasoning	54.3%	55%	high	—	External ↗

Source: Epoch AI Benchmarking Hub (CC BY 4.0). “External” rows are leaderboard results Epoch collects from third parties. Best reported setting per benchmark is shown.

API list price over time

Price history

$ per 1M tokens

13 observations (OpenRouter listing + Internet Archive snapshots). History accumulates with every ingest run; a single point means no change has been observed yet.