OpenAIUnited States· released Dec 2025

GPT-5.2

Name: GPT-5.2
Author: OpenAI

Ranked #19, 32.3% behind the frontier on average.

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

Third-party publishedMedium confidence (7+ comparable tests, 3+ Index categories)

Snapshot September 8, 2026

OpenCharts Index

67.7

Ranked #19 (17 to 25) of 36 · 32.3% behind the frontier on average · 21 comparable tests · 3 counting categories

Drop any single test and the Index lands between 63.0 and 72.0, anywhere from #17 to #25. Neighbours inside that range are ties.

8 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-5.2, in 21 results.

Strongest in agentic (78.6) and reasoning (70.5). It holds the best published result on GDPval. It trails the frontier most in math, 46.1% behind on average. No comparable result yet in long context. Between 30% and 35% behind the best published results, on average. Drop any single test and it would sit anywhere from #17 to #25.

Best published result onGDPval49.7%

Strongest category

Agentic

Long-horizon tasks with tools: terminals, desktops, browsers, whole jobs.

78.6

GDPval49.7% · best
METR Time Horizon5.9 h · 84% of best
Terminal-Bench64.9% · 77% of best
DeepResearch Bench41.1% · 74% of best
APEX-Agents34.4% · 73% of best
Vending-Bench 2$3,591 · 63% of best

Mean of 6 of the category's 8 comparable tests. Counts toward the Index.

Best single result

GDPval

Produce a real work deliverable across 44 occupations, judged against what industry professionals hand in.

49.7%

Best published result of 8 models.

Judged against professionals, so it measures economic value rather than trivia.

Where it trails

ProofBench

Write a complete, rigorous proof, graded for rigor rather than for a final answer.

15.0%

85.0% behind the best published result (100.0%) · #44 of 61.

A proof shows the reasoning itself. A right answer with a wrong argument scores nothing.

Not measured yet

No comparable result yet in this category. A blank is a blank, never a zero, and it does not lower the Index.

Long context

Covered, but not averaged into the Index:

Knowledgesingle-test category, shown beside the Index · 49.1
Coding3 of 7 tests · needs 4 · mean 62.1
Multimodalsingle-test category, shown beside the Index · 80.4
Human preferencesingle-test category, shown beside the Index · 80.3

Every score, with its source

See it at work

What GPT-5.2 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 80.4 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.

1,244

40% expected win rate against the board leader (1,313)
#37 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-5.2 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-5.2

#19·Index 67.7C+

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
70.5
99.6
Knowledge
49.1
100.0
Coding
62.1
97.7
Math
53.9
99.8
Agentic
78.6
98.5
Multimodal
80.4
—
Human preference
80.3
—

Shared tests · biggest gaps first

ProofBenchMath
15.0%
99.0%
−84.0
Code Arena (WebDev)Coding
1,416
1,797
−79.9
FrontierMath Tier 4Math
31.7%
97.6%
−67.5
SimpleQA VerifiedKnowledge
37.1%
75.6%
−50.9
ARC-AGI-2Reasoning
52.9%
95.0%
−44.3
FrontierMath (Tiers 1–3)Math
67.4%
93.7%
−28.1
APEX-AgentsAgentic
34.4%
46.7%
−25.9
WeirdMLCoding
72.2%
92.9%
−22.3
ARC-AGI-1Reasoning
86.2%
98.5%
−12.5
GPQA DiamondReasoning
91.4%
95.8%
−6.2
Mock AIME 2024–2025Math
96.1%
100.0%
−3.9

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

Reasoning70.5 · 5/6
Knowledgebeside49.1 · 1/1
Coding62.1 · 3/7
Math53.9 · 4/5
Agentic78.6 · 6/8
Multimodalbeside80.4 · 1/1
Human preferencebeside80.3 · 1/1
Long context—

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

91.4%

93.8% of the best (95.8%), above chance
77th percentile of 78 · n = 198

2025-12-11 · xhigh

GPT-5.2

GPT-5.2, in 21 results.

What GPT-5.2 was asked to do, and how it did.

GPT-5.2 against whoever you pick.

Every test, every source

Compare with its neighbours.

The best models, ranked here, working inside Theo.