OpenAIUnited States· released Aug 2025

GPT-5 mini

Name: GPT-5 mini
Author: OpenAI

Provisional: 19 comparable tests so far. Ranked once it reaches 6 across 3 counting categories.

OpenAI, United States. One of 26 models from this lab on the Index.

Third-party publishedLow confidence (thin coverage)

Snapshot September 8, 2026

OpenCharts Index

43.0

Provisional: not enough comparable tests or counting categories to rank yet · 19 comparable tests · 2 counting categories

Drop any single test and the Index lands between 37.0 and 48.4.

6 of 8 hardest tests · share of the frontier

Distance to today's frontier on the hardest tests. Not a measure of intelligence.

What it's known for

GPT-5 mini, in 19 results.

Strongest in math (51.5) and reasoning (34.5). It trails the frontier most in reasoning, 65.5% behind on average. Between 50% and 60% behind the best published results, on average.

Strongest category

Math

Competition and research mathematics, graded on the final answer or the proof.

51.5

MATH Level 597.8% · 100% of best
Mock AIME 2024–202586.7% · 87% of best
FrontierMath (Tiers 1–3)46.7% · 50% of best
FrontierMath Tier 412.2% · 13% of best
ProofBench9.0% · 9% of best

Mean of 5 of the category's 5 comparable tests. Counts toward the Index.

Best single result

MATH Level 5

Solve the hardest tier of the MATH competition dataset to an exact final answer.

97.8%

99.7% of the best (98.1%) · #2 of 6.

Multi-step algebra and geometry, graded only on the final answer.

Where it trails

Vending-Bench 2

Run a simulated vending business for a year: ordering, pricing and cash flow. The score is the final balance.

$-31

100.0% behind the best published result ($11,182) on a log scale · #44 of 53.

Long-horizon decisions whose consequences compound. It exposes models that drift or forget.

Not measured yet

Covered, but not averaged into the Index:

Knowledgesingle-test category, shown beside the Index · 28.6
Coding3 of 7 tests · needs 4 · mean 65.8
Agentic2 of 8 tests · needs 4 · mean 20.6
Multimodalsingle-test category, shown beside the Index · 64.0
Human preferencesingle-test category, shown beside the Index · 67.4
Long contextsingle-test category, shown beside the Index · 62.5

Every score, with its source

See it at work

What GPT-5 mini was asked to do, and how it did.

Every test it has taken, by category, with the kind of task it faced, who produced the number and where the result landed against the best published one. The examples are original and illustrative, never items from the datasets themselves.

Human preference · mean 67.4 across 1 comparable test · a single-test category, shown beside the Index and never averaged into it

Text ArenaLeaderboard

Answer the same prompt as a rival model; real users vote blind on which answer they prefer.

1,389

34% expected win rate against the board leader (1,507)
#87 of 92

high · board 2026-09-02

Source: Arena

Every result is read against the best published one on its test: percent scores as a share of the best above chance, Arena ratings as the expected win rate against the board leader, open-ended values on a log scale. Sitting a harder exam never lowers a model. Every number links to the publisher that produced it.

Head to head

GPT-5 mini against whoever you pick.

Choose a rival. Every counted test both have taken appears side by side as a share of the best published result on that test, with the category means above. A win is a higher share; a gap under 2 points is a tie, because one item on a 45-problem exam is 2.2 points and publishers' own standard errors are of that order.

GPT-5 mini

provisional·Index 43.0D

0 : 11

wins · 0 ties · 11 shared

GPT-6 Astra

#1·Index 99.0A+

GPT-6 Astra comes out ahead on 11 of the 11 tests both have taken.

Category means

Reasoning
34.5
99.6
Knowledge
28.6
100.0
Coding
65.8
97.7
Math
51.5
99.8
Agentic
20.6
98.5
Multimodal
64.0
—
Human preference
67.4
—
Long context
62.5
—

Shared tests · biggest gaps first

CritPtReasoning
0.0%
31.7%
−98.2
ARC-AGI-2Reasoning
4.4%
95.0%
−95.3
ProofBenchMath
9.0%
99.0%
−90.0
FrontierMath Tier 4Math
12.2%
97.6%
−87.5
SimpleQA VerifiedKnowledge
21.6%
75.6%
−71.4
FrontierMath (Tiers 1–3)Math
46.7%
93.7%
−50.2
ARC-AGI-1Reasoning
54.3%
98.5%
−44.8
WeirdMLCoding
52.7%
92.9%
−43.3
GPQA DiamondReasoning
75.0%
95.8%
−29.3
SciCodeCoding
39.2%
56.5%
−27.8
Mock AIME 2024–2025Math
86.7%
100.0%
−13.3

Bars are distances to the best published result on each test (percent scores above chance, Arena ratings as win rate against the leader, open-ended values on a log scale). Raw values as the publishers report them.

Open GPT-6 Astra

Scores

Every test, every source

The best published run per benchmark, the raw value as the publisher reports it, who produced the number, how close it comes to the best published result on that test, and where the model sits among every model scored on it. Percent scores are read above chance, Arena ratings as a win rate against the board leader, open-ended values on a log scale. A test with too few models to compare against is shown but not counted.

Category means

Reasoning34.5 · 5/6
Knowledgebeside28.6 · 1/1
Coding65.8 · 3/7
Math51.5 · 5/5
Agentic20.6 · 2/8
Multimodalbeside64.0 · 1/1
Human preferencebeside67.4 · 1/1
Long contextbeside62.5 · 1/1

Mean · tests taken of the category's comparable tests. A faded category either rests on fewer than half of them or is a single-test category (marked “beside”), which is shown next to the Index and never averaged into it.

GPQA DiamondEpoch-runReasoning

75.0%

70.7% of the best (95.8%), above chance
6th percentile of 78 · n = 198

2025-08-07 · high

GPT-5 mini

GPT-5 mini, in 19 results.

What GPT-5 mini was asked to do, and how it did.

GPT-5 mini against whoever you pick.

Every test, every source

Compare with its neighbours.

The best models, ranked here, working inside Theo.