VerdictPal
Join

Models / OpenAI

GPT-5.2

GPT-5 · Released 2025-12-11

OpenAI's December 2025 frontier. 410K context, 128K output, configurable reasoning effort. $1.75/$14 per 1M tokens.

Previous-gen GPT-5 for complex professional work. Superseded by GPT-5.4/5.5 but still available at lower cost.

#45 of 168 on AA Index · current snapshot
43 AA Index · 18 benchmark rows · 5 sourcesPosition is benchmark-specific — not a cross-family or cross-source ranking.
Composite69
Context410K
Input / 1M$1.8
Output / 1M$14
Knowledge cutoff2025-08-01
Statussolid

Family profile

Best published score in each covered benchmark family.

85/100 avg
Knowledge
87
Reasoning
90
Math
89
Coding
95
Agentic
72
Long context
79

6 tested benchmark families

Benchmark placements

Where GPT-5.2 places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.

BenchmarkFamilyRankScoreSourceDate
HumanEvalCoding#4/ 1595CodeSOTA HumanEval leaderboard2025-12-01
LiveCodeBenchCoding#5/ 3189Artificial Analysis2026-09-01
GDPvalAgentic#6/ 672OpenAI GDPval2025-12-11
FrontierMathMath#8/ 1541Epoch AI FrontierMath Tiers 1-3 (v2) CSV2026-08-15
AA-LCRLong context#10/ 17979Artificial Analysis2026-09-01
ARC-AGIReasoning#10/ 1453ARC Prize leaderboard2026-05-15
MMLU-ProKnowledge#10/ 3987Artificial Analysis2026-09-01
SWE-bench VerifiedCoding#11/ 2380SWE-bench official leaderboard2026-05-30
MATHMath#12/ 2587OpenAI model release notes2025-12-01
AIMEMath#18/ 3789OpenAI model release notes2025-12-01
IFBenchReasoning#27/ 12775Artificial Analysis2026-09-01
SciCodeCoding#31/ 18052Artificial Analysis2026-09-01
GPQA DiamondReasoning#41/ 18590Artificial Analysis2026-09-01
Humanity's Last ExamReasoning#43/ 18038Artificial Analysis2026-09-01
Artificial Analysis Intelligence IndexReasoning#56/ 18743Artificial Analysis2026-09-01
Terminal-BenchAgentic#93/ 18447Artificial Analysis2026-09-01
AA time to first tokenPerformance#131/ 186Artificial Analysis2026-09-01
AA output speedPerformance#132/ 187Artificial Analysis2026-09-01
Family context

The three highest-scoring models with pages in each capability family. Where GPT-5.2 shows up, it's highlighted.

KnowledgeGPT-5.2 · 87
  1. 1
    GPT 5.5
    91
  2. 2
    Claude Opus 4.8
    91
  3. 3
    Claude Sonnet 4.6
    91
ReasoningGPT-5.2 · 90
  1. 1
    Claude Mythos Preview
    95
  2. 2
    Gemini 3.7 Flash
    95
  3. 3
    Grok 4.6
    95
MathGPT-5.2 · 89
  1. 1
    OpenAI o3
    99
  2. 2
    GPT-5 (medium)
    99
  3. 3
    GPT-5 (high)
    99
CodingGPT-5.2 · 95
  1. 1
    GPT 5.5
    96
  2. 2
    Claude Opus 4.8
    96
  3. 3
    Claude Opus 4.6
    96
AgenticGPT-5.2 · 72
  1. 1
    GPT 5.5
    100
  2. 2
    Claude Opus 4.8
    100
  3. 3
    Claude Fable 5
    100
Long contextGPT-5.2 · 79
  1. 1
    Muse Spark 1.2
    83
  2. 2
    Kimi K3
    83
  3. 3
    Gemini 3.5 Flash
    81
Newest receipts

Benchmark rows added to the public ledger for GPT-5.2 in the last 120 days. Older rows live in the full table below.

14 recent
BenchmarkFamilyScoreSourceDays ago
AA-LCRLong context79.33%Artificial Analysis5
AA output speedPerformance0 t/sArtificial Analysis5
AA time to first tokenPerformance0.00sArtificial Analysis5
Artificial Analysis Intelligence IndexReasoning43.3%Artificial Analysis5
GPQA DiamondReasoning90.3%Artificial Analysis5
Humanity's Last ExamReasoning37.7%Artificial Analysis5
IFBenchReasoning75.44%Artificial Analysis5
LiveCodeBenchCoding88.9%Artificial Analysis5
MMLU-ProKnowledge87.4%Artificial Analysis5
SciCodeCoding52.1%Artificial Analysis5
Terminal-BenchAgentic46.97%Artificial Analysis5
FrontierMathMath40.7%Epoch AI FrontierMath Tiers 1-3 (v2) CSV22
SWE-bench VerifiedCoding80%SWE-bench official leaderboard99
ARC-AGIReasoning52.9%ARC Prize leaderboard114
Family coverage

A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.

Knowledge87
Reasoning90
Math89
Coding95
Agentic72
Long context79
Price vs. performance
0255075100$0.10$0.30$1$3$10$30input USD / 1M tokens · log scale →← intelligence

Hover any dot for name, score, and input price. Keyboard: tab through the top twelve, or use the ranking below.

AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMetaMiniMaxxAIZhipuXiaomiOtherMeituanCohereNVIDIA

Full benchmark ledger

Every catalog benchmark for GPT-5.2. Scores link to the original source; gaps mean no public row exists yet.

18 sourced rows

18 of 37 catalog benchmarks have a sourced row for GPT-5.2.

Knowledge

Reasoning

Math

Coding

Agentic

Long context

Performance

Changelog
  1. Released GPT-5.2 — 410K context, configurable reasoning, $1.75/$14.
Sources