Enter opens the highlighted result. Esc closes.

GPT-5.2-Codex vs GPT-6.2 Sol

GPT-6.2 Sol leads GPT-5.2-Codex on coding, 83.1 to 45.4 on the Vibe Coding Index, and GPT-6.2 Sol costs 1.4x less per output token ($10.00 vs $14.00 per million). Updated November 22, 2026; 4 benchmarks have a result for both.

GPT-5.2-CodexOpenAI
Coding
45.4
Overall
No data
Output $/M
$14.00
0
GPT-6.2 SolOpenAI
Coding
83.1
Overall
79.9
Output $/M
$10.00
4
vs boards won of 4 shared

GPT-5.2-Codex or GPT-6.2 Sol: which to pick

Pick GPT-6.2 Sol for coding: it scores 83.1 on the Vibe Coding Index against 45.4 for GPT-5.2-Codex.

GPT-5.2-Codex is neither cheaper nor larger in context, and leads no benchmark both models were measured on.

Shared results
4
GPT-5.2-Codex only
18
GPT-6.2 Sol only
52
Like-for-like categories
1 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
GPT-5.2-Codex 45.4 from 3 benchmarks, GPT-6.2 Sol 83.1 from 4 benchmarks.
Widest category
Coding, 49.1 points apart on average over 4 shared benchmarks.
Widest single gap
LMArena WebDev: GPT-5.2-Codex 1339 against GPT-6.2 Sol 1754, 59.3 points apart on a 0 to 100 scale.
Output price
GPT-5.2-Codex $14.00 against GPT-6.2 Sol $10.00 per million output tokens, 1.4x apart.
Context window
GPT-5.2-Codex takes 400K tokens, GPT-6.2 Sol 1.05M.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryGPT-5.2-CodexBarsGPT-6.2 SolTies
Coding04
All shared boards04
CodingLike-for-likeSix of these boards feed the Vibe Coding Index04
Coding benchmarks for GPT-5.2-Codex and GPT-6.2 Sol
BenchmarkGPT-5.2-CodexGPT-6.2 SolLeader
SWE-bench Verified72.8%Not measuredOnly GPT-5.2-Codex measured
SWE-bench Multilingual66.3%Not measuredOnly GPT-5.2-Codex measured
LiveBench codingNot measured83.7%Only GPT-6.2 Sol measured
LMArena WebDev13391754GPT-6.2 Sol
LiveCodeBench (Vals AI)88.0%Not measuredOnly GPT-5.2-Codex measured
SWE-bench Verified (Vals AI)72.4%Not measuredOnly GPT-5.2-Codex measured
Terminal-Bench 4 (Vals AI)Not measured60.6%Only GPT-6.2 Sol measured
Vibe Code Bench (Vals AI)37.9%87.6%GPT-6.2 Sol
Terminal-Bench Hard (Artificial Analysis run)37.1%Not measuredOnly GPT-5.2-Codex measured
Terminal-Bench 4.0 (Artificial Analysis run)Not measured60.1%Only GPT-6.2 Sol measured
FrontierCode 2.0Not measured44.4%Only GPT-6.2 Sol measured
Terminal-Bench66.5%Not measuredOnly GPT-5.2-Codex measured
ALE-Bench1299.9Not measuredOnly GPT-5.2-Codex measured
LMArena Text: codingNot measured1440Only GPT-6.2 Sol measured
LMArena Text: software and ITNot measured1434Only GPT-6.2 Sol measured
LMArena WebDev: React13291761GPT-6.2 Sol
LMArena WebDev: HTML13481840GPT-6.2 Sol
GameBench 2Not measured94.4Only GPT-6.2 Sol measured
Terminal-Bench 2.066.5%Not measuredOnly GPT-5.2-Codex measured
Neither model has a result on the other 33 boards here
IntelligenceDirectional onlyPublic boards, shown as published, kept out of the index1 board, none shared
Intelligence benchmarks for GPT-5.2-Codex and GPT-6.2 Sol
BenchmarkGPT-5.2-CodexGPT-6.2 SolLeader
AA Intelligence IndexNot measured57.3Only GPT-6.2 Sol measured
Neither model has a result on the other 1 board here
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index17 boards, none shared
Agentic work benchmarks for GPT-5.2-Codex and GPT-6.2 Sol
BenchmarkGPT-5.2-CodexGPT-6.2 SolLeader
Design Arena: agents building web apps1019Not measuredOnly GPT-5.2-Codex measured
Design Arena: agents building full-stack apps971Not measuredOnly GPT-5.2-Codex measured
Design Arena: agents building mobile apps1082Not measuredOnly GPT-5.2-Codex measured
Design Arena: agents building Android apps1176Not measuredOnly GPT-5.2-Codex measured
Design Arena: agents building Godot games1142Not measuredOnly GPT-5.2-Codex measured
LiveBench researchNot measured90.1%Only GPT-6.2 Sol measured
LiveBench browserNot measured88.2%Only GPT-6.2 Sol measured
LiveBench long agenticNot measured63.6%Only GPT-6.2 Sol measured
Tax Agent Bench (Vals AI)Not measured24.8%Only GPT-6.2 Sol measured
Legal Research Bench (Vals AI)Not measured37.0%Only GPT-6.2 Sol measured
Harvey's Legal Agent Benchmark (Vals AI)Not measured12.1%Only GPT-6.2 Sol measured
BioMysteryBench (Vals AI)Not measured77.8%Only GPT-6.2 Sol measured
Terminal-Bench-Science (Vals AI)Not measured58.6%Only GPT-6.2 Sol measured
Vals Science IndexNot measured57.8%Only GPT-6.2 Sol measured
τ²-Bench Telecom92.1%Not measuredOnly GPT-5.2-Codex measured
APEX-AgentsNot measured66.6%Only GPT-6.2 Sol measured
LMArena SearchNot measured1179Only GPT-6.2 Sol measured
Neither model has a result on the other 27 boards here
ReasoningDirectional onlyPublic boards, shown as published, kept out of the index13 boards, none shared
Reasoning benchmarks for GPT-5.2-Codex and GPT-6.2 Sol
BenchmarkGPT-5.2-CodexGPT-6.2 SolLeader
LiveBench reasoningNot measured90.6%Only GPT-6.2 Sol measured
MysteryMechanism (Vals AI)Not measured36.9%Only GPT-6.2 Sol measured
GPQA Diamond (Artificial Analysis run)89.9%Not measuredOnly GPT-5.2-Codex measured
Humanity's Last Exam (Artificial Analysis run)35.7%Not measuredOnly GPT-5.2-Codex measured
HLE-Diamond (Artificial Analysis run)Not measured64.0%Only GPT-6.2 Sol measured
CritPt v2, hard subset (Artificial Analysis run)Not measured64.0%Only GPT-6.2 Sol measured
ARC-AGINot measured97.5%Only GPT-6.2 Sol measured
ARC-AGI-2Not measured85.8%Only GPT-6.2 Sol measured
SimpleBenchNot measured69.9%Only GPT-6.2 Sol measured
LMArena Text: hard promptsNot measured1419Only GPT-6.2 Sol measured
LMArena Text: hard prompts in EnglishNot measured1413Only GPT-6.2 Sol measured
Kagi LLM BenchmarkNot measured67.0%Only GPT-6.2 Sol measured
Humanity's Last Exam (Scale run)Not measured56.5%Only GPT-6.2 Sol measured
Neither model has a result on the other 13 boards here
MathDirectional onlyPublic boards, shown as published, kept out of the index1 board, none shared
Math benchmarks for GPT-5.2-Codex and GPT-6.2 Sol
BenchmarkGPT-5.2-CodexGPT-6.2 SolLeader
MathArenaNot measured91.2%Only GPT-6.2 Sol measured
Neither model has a result on the other 13 boards here
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index3 boards, none shared
Knowledge benchmarks for GPT-5.2-Codex and GPT-6.2 Sol
BenchmarkGPT-5.2-CodexGPT-6.2 SolLeader
AA-OmniscienceNot measured40.4Only GPT-6.2 Sol measured
LMArena Text: scienceNot measured1432Only GPT-6.2 Sol measured
LMArena Text: business and financeNot measured1383Only GPT-6.2 Sol measured
Neither model has a result on the other 18 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index4 boards, none shared
Web design benchmarks for GPT-5.2-Codex and GPT-6.2 Sol
BenchmarkGPT-5.2-CodexGPT-6.2 SolLeader
Design Arena: game devNot measured1437Only GPT-6.2 Sol measured
Design Arena: 3DNot measured1482Only GPT-6.2 Sol measured
Design Arena: all code categoriesNot measured1405Only GPT-6.2 Sol measured
LiveBench designNot measured80.0%Only GPT-6.2 Sol measured
Neither model has a result on the other 5 boards here
GeneralDirectional onlyPublic boards, shown as published, kept out of the index16 boards, none shared
General benchmarks for GPT-5.2-Codex and GPT-6.2 Sol
BenchmarkGPT-5.2-CodexGPT-6.2 SolLeader
Vals IndexNot measured66.1%Only GPT-6.2 Sol measured
LiveBench data analysisNot measured81.1%Only GPT-6.2 Sol measured
IFBench77.6%Not measuredOnly GPT-5.2-Codex measured
Long context reasoning (AA-LCR)82.3%Not measuredOnly GPT-5.2-Codex measured
LMArena TextNot measured1414Only GPT-6.2 Sol measured
LMArena Text: instruction followingNot measured1434Only GPT-6.2 Sol measured
LMArena Text: creative writingNot measured1403Only GPT-6.2 Sol measured
LMArena Text: EnglishNot measured1419Only GPT-6.2 Sol measured
LMArena Text: non-EnglishNot measured1405Only GPT-6.2 Sol measured
LMArena Text: ChineseNot measured1434Only GPT-6.2 Sol measured
LMArena Text: RussianNot measured1401Only GPT-6.2 Sol measured
LMArena Text: multi-turnNot measured1420Only GPT-6.2 Sol measured
LMArena Text: longer queriesNot measured1427Only GPT-6.2 Sol measured
LMArena Text: writing and languageNot measured1420Only GPT-6.2 Sol measured
LMArena Text: entertainment and mediaNot measured1408Only GPT-6.2 Sol measured
MultiChallengeNot measured72.0%Only GPT-6.2 Sol measured
Neither model has a result on the other 33 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-5.2-Codex and GPT-6.2 Sol
BenchmarkGPT-5.2-CodexGPT-6.2 SolLeader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

GPT-5.2-Codex$0.0088Fits in one request
GPT-6.2 Sol$0.0070Fits in one request

GPT-6.2 Sol costs less

Repository review

50K input + 3K output tokens

GPT-5.2-Codex$0.130Fits in one request
GPT-6.2 Sol$0.130Fits in one request

GPT-5.2-Codex costs less

Agent loop

200K input + 10K output tokens

GPT-5.2-Codex$0.490Fits in one request
GPT-6.2 Sol$0.500Fits in one request

GPT-5.2-Codex costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of GPT-5.2-Codex and GPT-6.2 Sol
SpecificationOpenAI GPT-5.2-CodexOpenAI GPT-6.2 Sol
MakerOpenAIOpenAI
WeightsNo public weights listedNo public weights listed
ReleasedJanuary 14, 2026November 17, 2026
Context window400K1.05M
Input $/Mtok$1.75$2.00
Output $/Mtok$14.00$10.00
Coding index45.483.1
Independent measurements behind it34
Cost per solved task$0.43Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5.2-Codex better than GPT-6.2 Sol for coding?

Pick GPT-6.2 Sol for coding: it scores 83.1 on the Vibe Coding Index against 45.4 for GPT-5.2-Codex. GPT-5.2-Codex is neither cheaper nor larger in context, and leads no benchmark both models were measured on.

Which is cheaper, GPT-5.2-Codex or GPT-6.2 Sol?

GPT-5.2-Codex costs $1.75 per million input tokens and $14.00 per million output tokens; GPT-6.2 Sol costs $2.00 and $10.00. Cost per solved task is known for GPT-5.2-Codex only.

Which has the bigger context window, GPT-5.2-Codex or GPT-6.2 Sol?

GPT-6.2 Sol, with 1.05M tokens against 400K.

How many benchmarks is the GPT-5.2-Codex vs GPT-6.2 Sol comparison based on?

4 benchmarks have a result for both models as of November 22, 2026. GPT-5.2-Codex's coding index rests on 3 independent benchmarks and GPT-6.2 Sol's on 4. Where only one model has a result, the page says so instead of naming a winner.

Compare next

Embed this comparison on your site

Free under CC BY 4.0, while the source link stays visible. Preview the frame · Terms and other embeds

<iframe src="https://llmbenchmark.ai/embed/compare/gpt-5-2-codex-vs-gpt-6-2-sol" title="GPT-5.2-Codex vs GPT-6.2 Sol, LLMBenchmark.ai" width="100%" height="360" style="border:0;max-width:100%" loading="lazy"></iframe>
<p style="font:13px/1.4 sans-serif;margin:6px 0 0">Source: <a href="https://llmbenchmark.ai/compare/gpt-5-2-codex-vs-gpt-6-2-sol">LLMBenchmark.ai, GPT-5.2-Codex vs GPT-6.2 Sol</a></p>