OpenAIReleased Nov 17, 2026
NewProprietary

gpt-6.2-sol

Strongest in Mathematical Reasoning (#1 of 186) and Computer Use (#5 of 186); weakest in Reliability (#41 of 188). Three of seven dimensions are estimated, because no results are published yet.

Composite IQ ?
140
#6 of 186 ranked models · ahead of 97%
OpenAI's #4 model, behind gpt-6.2-astra (142)
Effective cost
$6.41/1M
$2 in · $10 out
Speed
Not measured
No speed data yet
Context
1Mtokens
1,050,000 tokens in
Evidence
4of 7 dims
6 of 39 benchmarks

IQ profile

Seven dimensions, averaged equally into Composite IQ
Compare shape with
100 avg human130160● Abstract130#8● Math157#1● Academic135#10 · est.● Coding153#4 · est.● Computer Use150#5● Reliability122#41● Knowledge Work132#13 · est.
MeasuredEstimated95% range
#1 in dimensionTop-10 median

IQ dimensions

Ranks are among models scored in each dimension. Ranges are model-based estimates, not calibrated confidence intervals.

DimensionIQPosition · 70–160RankEvidencevs Top 10
130measured
#8 / 221
3/5
-2.5
157measured
#1 / 186
1/5
+7.5
135estimated
#10 / 223
0/6
0
153estimated
#4 / 187
0/6
+2
150measured
#5 / 186
1/5
+3.5
122measured
#41 / 188
1/6
-13.5
132estimated
#13 / 186
0/6
0
gpt-6.2-sol95% range (model-based, uncalibrated)Top-10 median#1 in dimension100 = average human

Benchmark results

6 of 39 Composite IQ benchmarks have a published result for gpt-6.2-sol. Benchmarks without a result stay listed so gaps are visible.

Abstract Reasoning IQ 130 · measuredBest: #5 of 29 on ARC-AGI-33/5
BenchmarkScorePercentileRankLeader · source
ARC-AGI-3
Full source leaderboard reconciliation · Nov 18
54.9%
#5 / 29
Leader fable-5.2 · 81.2%
Full source leaderboard reconciliation · Nov 18
ARC-AGI-1
Full source leaderboard reconciliation · Nov 18
97.5%
#13 / 73
Leader fable-5 · 98.5%
Full source leaderboard reconciliation · Nov 18
ARC-AGI-2
Full source leaderboard reconciliation · Nov 18
85.8%
#14 / 73
Leader gpt-6.2-astra · 95.8%
Full source leaderboard reconciliation · Nov 18
Not yet reported
Chess PuzzlesMystery Game Puzzles
Mathematical Reasoning IQ 157 · measuredBest: #2 of 70 on MathArena1/5
BenchmarkScorePercentileRankLeader · source
MathArena
Full source leaderboard reconciliation · Nov 18
89.5%
#2 / 70
Leader gpt-6.2-astra · 92.2%
Full source leaderboard reconciliation · Nov 18
Not yet reported
FrontierMath Tier 4FrontierMath Tier 1-3ProofBench 1.1AIME
Academic Reasoning IQ 135 · estimatedNo published results0/6
No published results yet for Academic Reasoning. The 135 shown is estimated from gpt-6.2-sol's other dimensions and will be replaced as results come in. How estimates work
Humanity's Last ExamGPQA DiamondCritPtSciCodeMMLU-ProMMMU-Pro
Programmatic Reasoning IQ 153 · estimatedNo published results0/6
No published results yet for Programmatic Reasoning. The 153 shown is estimated from gpt-6.2-sol's other dimensions and will be replaced as results come in. How estimates work
LiveCodeBenchIOI V2Terminal-Bench 4.0ProgramBenchSWE-rebenchMirrorCode
Computer Use IQ 150 · measuredBest: #11 of 77 on Arena.ai Agent Arena1/5
BenchmarkScorePercentileRankLeader · source
Arena.ai Agent Arena
Arena.ai Agent Arena · Nov 19
11.3
#11 / 77
Leader fable-5.2 · 27.4
Arena.ai Agent Arena · Nov 19
Not yet reported
BrowseCompOSWorld-VerifiedMCP AtlasAgents' Last Exam
Reliability IQ 122 · measuredBest: #52 of 96 on BullshitBench v21/6
BenchmarkScorePercentileRankLeader · source
BullshitBench v2
Full source leaderboard reconciliation · Nov 19
42.0%
#52 / 96
Leader opus-4.8 · 94.0%
Full source leaderboard reconciliation · Nov 19
Not yet reported
SimpleQA VerifiedaaOmniscienceIFBenchAA Long Context Reasoning v1.1FACTS Grounding
Knowledge Work IQ 132 · estimatedNo published results0/6
No published results yet for Knowledge Work. The 132 shown is estimated from gpt-6.2-sol's other dimensions and will be replaced as results come in. How estimates work
AA-Briefcase v1.1GDPval-AA v2.1AutomationBench-AAGDP.pdfAA-AnalystAgentEnterpriseOps-Gym-AA

Release & availability

Facts from the publisher's announcement and documentation.

Released
Developer
OpenAI
Access
Available as gpt-6.2-sol through the OpenAI API and in ChatGPT Work and Codex for eligible Plus, Pro, Enterprise, and Edu users; Enterprise and Edu access depends on workspace settings.
Weights
ProprietaryHosted access only
Inputs
textimage
Context
1,050,000 tokens
API price
$2 in · $10 outper 1M tokens · full pricing

What OpenAI says

  • OpenAI announced GPT-6.2 Sol on November 17, 2026 for coding and everyday agent workflows.
  • The API model supports a 1,050,000-token context window and reasoning effort.
  • Standard short-context API rates are $2 per million input tokens and $10 per million output tokens.

Limitations

  • Audio and video inputs are not supported.
  • The API rejects reasoning effort none.
  • Requests above 272K input tokens incur higher rates for the full request.

Publisher statements, not AI IQ measurements.

About this score

How much of gpt-6.2-sol's IQ is measured, and how the rest is estimated.

6/ 39benchmark results
4/ 7dimensions measured
3estimated
3/5Abstract
1/5Math
0/6Academic
0/6Coding
1/5Computer Use
1/6Reliability
0/6Knowledge Work
Bar height = share of benchmarks with resultsEstimatedNo score
Measured results are shown as published

Every score in the benchmark tables comes from its source unchanged. Composite IQ averages all seven dimensions equally.

Estimates fill the gaps

Academic Reasoning, Programmatic Reasoning and Knowledge Work have no published results yet, so their IQ is estimated from gpt-6.2-sol's other dimensions. Missing benchmarks can also draw on its predecessor, GPT 6.1 Sol; a predecessor is an estimation input, not evidence that this model has the predecessor's measured results.

A missing result is not a zero

Benchmark versions and evaluation harnesses can differ between sources, so small gaps between models may not be meaningful.

Read the scoring methodology →

Compare gpt-6.2-sol

Side-by-side pages for the models it is most often weighed against.

gpt-6.2-solComparison model
Browse all comparisons →

Frequently asked questions

Quick answers about gpt-6.2-sol, drawn from the same data as the rest of this page.

How AI IQ scores models →Get this data from the API →
What is the IQ of gpt-6.2-sol?
gpt-6.2-sol has an estimated IQ of 140, which ranks #6 out of 186 public models on the AI IQ leaderboard. The score combines seven equally weighted reasoning dimensions using measured benchmark results and conservative estimates for missing coverage.
What is gpt-6.2-sol best at?
Among the dimensions with published benchmark results, gpt-6.2-sol ranks highest in Mathematical Reasoning (IQ 157, #1 of 186) and Computer Use (IQ 150, #5 of 186). Its lowest-ranked measured dimension is Reliability (IQ 122, #41 of 188).
What is the EQ of gpt-6.2-sol?
gpt-6.2-sol doesn't have an EQ score yet. EQ (emotional reasoning) is calculated from EQ-Bench 4 and AttuneBench, and neither has published a result for gpt-6.2-sol. AI IQ doesn't estimate EQ from other benchmarks, so the score appears once one of those results is published.
Which benchmarks has gpt-6.2-sol been tested on?
gpt-6.2-sol has published results on 6 of the 39 benchmarks behind AI IQ's Composite IQ, including MathArena, Arena.ai Agent Arena, ARC-AGI-3, ARC-AGI-1 and ARC-AGI-2.
How much does gpt-6.2-sol cost?
gpt-6.2-sol costs $2.00 per million input tokens and $10.00 per million output tokens. Adjusted for how many tokens it actually consumes on real tasks, its effective cost works out to $6.41 per million input-plus-output tokens.
Is gpt-6.2-sol open source?
No — gpt-6.2-sol is a proprietary model. Its weights are not publicly available, so it is used through hosted APIs and products.
Who developed gpt-6.2-sol?
gpt-6.2-sol was developed by OpenAI, based in the United States.
When was gpt-6.2-sol released?
gpt-6.2-sol was publicly released on November 17, 2026.
How large is the context window of gpt-6.2-sol?
gpt-6.2-sol supports a context window of 1,050,000 tokens — the amount of text it can consider in a single request.
How fast is gpt-6.2-sol?
AI IQ doesn't have independent speed measurements for gpt-6.2-sol yet. Speed comes from independent tests of a model's hosted API, which aren't available for every model — especially new, restricted-access, or self-hosted ones.