OpenAIUnited States· released Aug 2025

GPT-5 nano

Name: GPT-5 nano
Author: OpenAI

Provisional: 14 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

35.8

Provisional: not enough comparable tests or counting categories to rank yet · 14 comparable tests · 2 counting categories

Drop any single test and the Index lands between 29.0 and 42.8.

4 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 nano, in 14 results.

Strongest in math (42.8) and reasoning (28.8). It trails the frontier most in reasoning, 71.2% behind on average. Below 40% of the best published results, on average.

Strongest category

Math

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

42.8

MATH Level 595.2% · 97% of best
Mock AIME 2024–202581.1% · 81% of best
FrontierMath (Tiers 1–3)20.0% · 21% of best
ProofBench12.0% · 12% of best
FrontierMath Tier 42.4% · 3% 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.

95.2%

97.1% of the best (98.1%) · #6 of 6.

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

Where it trails

FrontierMath Tier 4

Solve the hardest FrontierMath tier: problems that take expert mathematicians days.

2.4%

97.5% behind the best published result (97.6%) · #46 of 51.

The deepest end of the set. Progress here signals genuinely new capability.

Not measured yet

Covered, but not averaged into the Index:

Knowledgesingle-test category, shown beside the Index · 15.5
Coding1 of 7 tests · needs 4 · mean 41.0
Agentic1 of 8 tests · needs 4 · mean 25.7
Multimodalsingle-test category, shown beside the Index · 55.1
Human preferencesingle-test category, shown beside the Index · 54.6
Long contextsingle-test category, shown beside the Index · 21.9

Every score, with its source

See it at work

What GPT-5 nano 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.

Multimodal · mean 55.1 across 1 comparable test · a single-test category, shown beside the Index and never averaged into it

Vision ArenaLeaderboard

Answer prompts that include an image; real users vote blind between two models' answers.

1,145

28% expected win rate against the board leader (1,313)
#54 of 54

high · board 2026-08-27

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 nano 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 nano

provisional·Index 35.8F

0 : 9

wins · 0 ties · 9 shared

GPT-6 Astra

#1·Index 99.0A+

GPT-6 Astra comes out ahead on 9 of the 9 tests both have taken.

Category means

Reasoning
28.8
99.6
Knowledge
15.5
100.0
Coding
41.0
97.7
Math
42.8
99.8
Agentic
25.7
98.5
Multimodal
55.1
—
Human preference
54.6
—
Long context
21.9
—

Shared tests · biggest gaps first

FrontierMath Tier 4Math
2.4%
97.6%
−97.5
ARC-AGI-2Reasoning
2.6%
95.0%
−97.3
ProofBenchMath
12.0%
99.0%
−87.0
SimpleQA VerifiedKnowledge
11.7%
75.6%
−84.5
ARC-AGI-1Reasoning
20.7%
98.5%
−79.0
FrontierMath (Tiers 1–3)Math
20.0%
93.7%
−78.7
WeirdMLCoding
38.1%
92.9%
−59.0
GPQA DiamondReasoning
69.4%
95.8%
−37.2
Mock AIME 2024–2025Math
81.1%
100.0%
−18.9

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

Reasoning28.8 · 3/6
Knowledgebeside15.5 · 1/1
Coding41.0 · 1/7
Math42.8 · 5/5
Agentic25.7 · 1/8
Multimodalbeside55.1 · 1/1
Human preferencebeside54.6 · 1/1
Long contextbeside21.9 · 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

69.4%

62.8% of the best (95.8%), above chance
1th percentile of 78 · n = 198

2025-08-07 · high

Source: Epoch AI — AI Benchmarking Hub

ARC-AGI-2LeaderboardReasoning

2.6%

2.7% of the best (95.0%)
0th percentile of 51 · n = 120

2025-08-07 · high

Source: Epoch AI — AI Benchmarking Hub

ARC-AGI-1LeaderboardReasoning

20.7%