Tabbit
ResourcesBlogModels
Tabbit LogoTabbit

Tabbit — The AI Browser that Works for You

Topics

  • Best Browsers 2026: Rankings and Top Lists
  • AI Browser Guides and Resources
  • Browser Comparisons: Head-to-Head Guides
  • Browser Reviews and Explainers
  • Browser Alternatives
  • Browser Directory: Guides by Browser
  • Agentic Browser Resources
  • Browser Downloads and Install Guides
  • Browser Productivity Resources
  • SillyTavern Model Guides and Presets

Popular Guides

  • Best Browsers 2026: Top 10 Ranked & Compared
  • Best Free AI Browsers 2026: Plans & Prices Compared
  • Rabbit Browser: What It Is & Why Tabbit Is Better
  • Firefox vs Chrome 2026: Privacy, RAM & Speed Compared
  • Best Lightweight Browsers 2026: 7 Tested for RAM & Speed
  • AI Browser List 2026: 10 Browsers Compared
  • Neobrowser: Norton Neo, Download & Alternatives
  • Edge vs Chrome 2026: Speed, RAM & Battery Tested
  • Best Internet Browser for Android 2026
  • Best AI Browser 2026: Top 9 Tested & Ranked

Events

  • Tabbit Skill Competition
  • KPOP SBTI Fandom Personality Test
  • Tabbit Campus Creator Program
  • fifi's Picks: AI Skills for Research Papers
  • User Survey

About

  • Tabbit Blog
  • Press & Media
English
简体中文English
Review
MediaGPT-6.2 Sol

GPT-6.2 Sol: AI IQ Subscores, Benchmark Results, and Estimation Limits

Original source

AI IQ

AuthorLiberated Software LLC (credited on the website)

Tabbit curation2026-11-19

Read original

Takeaway

AI IQ gives GPT-6.2 Sol a Composite IQ of 140, ranking it 6th among 185 models ranked at the time. This score was not calculated from direct results across all seven dimensions: only mathematics, abstract reasoning, and reliability have direct benchmark results, while AI IQ estimates the other four dimensions. The reported raw results cover 5 of 39 benchmarks.

When to Use This

  • Useful for: Understanding how AI IQ aggregates GPT-6.2 Sol's published benchmark results, subscores, rankings, and estimated coverage.

  • Not suitable for: Inferring actual IQ, general capability, or performance on a specific business task from the Composite IQ. Treating the page's ranking as a comprehensive comparison under a shared test setup.

  • Model version covered: GPT-6.2 Sol, model ID gpt-6.2-sol.

  • Test environment or client: The page does not specify the reasoning levels, parameters, API snapshots, or execution clients used for this model's benchmarks.

  • Reasoning level and parameters: Not specified.

Methodology

AI IQ assigns 39 benchmarks to seven equally weighted dimensions: abstract reasoning, mathematical reasoning, academic reasoning, procedural reasoning, computer use, reliability, and knowledge work. The page reports each raw benchmark result and maps benchmark scores to IQ calibration anchors of 70, 85, 100, 115, 130, 145, and 160. Composite IQ is the arithmetic mean of the seven dimension scores, rounded to the nearest integer.

Only 3 of the model's 7 dimensions have direct benchmark results. The page labels academic reasoning, procedural reasoning, computer use, and knowledge work as estimated. Overall coverage is 5 of 39 benchmarks. AI IQ's methodology page explains that missing results are filled in with dimension estimates, so Composite IQ is not an average of only the tested dimensions. The site's methodology page provides supplementary explanation of the scoring. The model's scores and rankings in this article come only from the model profile page above.

Key Results

DimensionIQ scoreRankEvidence coverage
Abstract reasoning1308 / 220Measured, results for 3/5 benchmarks
Mathematical reasoning1571 / 185Measured, results for 1/5 benchmarks
Academic reasoning13510 / 222Estimated, 0/6
Procedural reasoning1534 / 186Estimated, 0/6
Computer use1474 / 185Estimated, 0/5
Reliability12340 / 187Measured, 1/6
Knowledge work13213 / 185Estimated, 0/6

The page reports a Composite IQ of 140, ranked 6 / 185, and labels it “ahead of 97%.” It lists GPT-6.1 Sol at 141. This score difference compares the site's composite scores and does not represent a performance gap on any single shared test.

Raw Data

The following are the five raw results shown on the profile page. The denominator in parentheses is the number of models in that benchmark's ranking list, not the number of test questions or samples.

BenchmarkGPT-6.2 Sol scorePage rankingLeader listed on the page
ARC-AGI-354.9%5 / 29fable-5.2, 81.2%
ARC-AGI-197.5%13 / 73fable-5, 98.5%
ARC-AGI-285.8%14 / 73gpt-6.2-astra, 95.8%
MathArena89.5%2 / 70gpt-6.2-astra, 92.2%
BullshitBench v242.0%51 / 95opus-4.8, 94.0%

The page labels each result “Full source leaderboard reconciliation · Nov 18,” but does not specify the year. It does not publish the number of questions, run count, sampling strategy, evaluation harness, prompts, reasoning parameters, or per-run raw records for each benchmark. Sample sizes cannot be inferred from ranking-list denominators. The page provides no speed measurements for the model.

Conclusions and Limitations

  • Data reported by the source: The five raw results, dimension IQ scores, composite score, and rankings are shown on the AI IQ profile page. They are reported values from a third-party aggregation page, not official OpenAI benchmark results or independent retests conducted for this article.

  • Keep measured and estimated results distinct: The four estimated dimensions shown on the page are not direct GPT-6.2 Sol benchmark results for those dimensions. Composite IQ includes estimated dimensions in its seven-dimension average, so 140 should not be read as a direct conversion of the five raw scores.

  • Limits of the IQ scale: AI IQ's methodology page says its mapping is provisional and relative to the site's own benchmarks, not a new validation study of human IQ. The page's IQ scores should not be interpreted as human IQ in the psychometric sense.

  • Limits of the interval: The 95% ranges shown in the profile chart are model-estimated intervals. The methodology page says they are not empirically calibrated confidence intervals and do not include all sources of uncertainty, such as the calibration ladder, task correlations, and source selection.

  • Limits of comparison: The denominators for subscore rankings differ, indicating different numbers of models ranked in each dimension. Benchmark sets, tools, and evaluation configurations may also differ. The profile page does not provide enough information to verify that models were evaluated with the same setup.

  • Dynamic snapshot: Rankings, total model counts, and leaders reflect what the page showed at the time of collection and should not be treated as fixed rankings.

Reproduction Notes

The profile page can be used to check the five published scores, the seven dimension estimate labels, and the rankings. The model's benchmark results or Composite IQ cannot be recalculated from it because the page does not fully publish model execution parameters, sample sizes, harness versions, or item-level data. The estimates also depend on AI IQ's calibration and missing-result handling rules. For an explanation of the scoring, see the same site's AI IQ methodology page (a supplementary methodology source, not a second data source for this model's results).

Curated by Tabbit

This is a third-party source navigator. Model versions, test environments, and personal experience vary; consult the original source.

GPT-6.2 Sol

Use and compare models in Tabbit

GPT-6.2 Sol

Related reviews

OfficialOpenAI official blog2026-11-17

GPT-6.2 Sol Launch Benchmarks and Cost Basis

OfficialGoogle DeepMind2026-11-18

GPT-6.2 Sol in Google DeepMind's Gemini 4 Neon Model Card Benchmark Table

OfficialArtificial Analysis2026-11-17

GPT-6.2 Sol Scores 3 Points Lower Than Its Predecessor on the Codex Coding Agent Index

MediaArtificial Analysis

Artificial Analysis: GPT-6.2 Sol Reasoning Levels, Task Time, and Pricing

GPT-6.2 Sol

Related showcases & prompts

OfficialOpenAI API documentation

GPT-6.2 Sol Model ID and Responses API Reasoning Configuration

OfficialOpenAI API documentation

GPT-6.2 Sol API Migration and Tool Configuration

OfficialChatGPT Learn (OpenAI)2026-11-17

Task Selection and Short Workflows for GPT-6.2 Sol in Codex

OfficialOpenAI API documentation

Review PRs in Parallel with GPT-6.2 Sol in the Responses API