OpenAI · proprietary

GPT-5.2

GPT-5.2 by OpenAI appears in 2 sources with Reasoning at 53.54. Best read for Code quality, Coding, LLM.

CreatorOpenAI

Release date2025-12-11

Knowledge cutoffNot published

Context400K tokens

Input price$1.75/M tokens

Output price$14/M tokens

Modalitytext + vision

CountryUS

Metrics

All source-backed metrics

Open in leaderboard

LLM Stats Rank13ranking · source_rank

Reasoning53.54reasoning · index_reasoning

Math50.09math · index_math

Coding34.61coding · index_code

Research26.09research · index_search

Writing33.06writing · index_communication

Vision34.73multimodal · index_vision

Tool calling27.73tool_calling · index_tool_calling

Healthcare43.89domain · index_healthcare

GPQA92.4 %reasoning · gpqa_score

AIME 2025100 %math · aime_2025_score

SWE-bench Verified80 %coding · swe_bench_verified_score

Code Arena1,519.42 %coding · coding_arena_score

Humanity Last Exam34.5 %reasoning · hle_score

ARC-AGI v252.9 %reasoning · arc_agi_v2_score

MMMLU89.6 %reasoning · mmmlu_score

MMMU-Pro79.5 %multimodal · mmmu_pro_score

BrowseComp65.8 %research · browsecomp_score

Toolathlon46.3 %tool_calling · toolathlon_score

MCP Atlas60.6 %tool_calling · mcp_atlas_score

FrontierMath40.3 %math · frontiermath_score

CharXiv-R82.1 %multimodal · charxiv_r_score

ScreenSpot Pro86.3 %multimodal · screenspot_pro_score

Context400,000 tokenscontext · context

Speed181.45 c/sperformance · throughput

Latency25,092 msperformance · latency

Input price1.75 $/Mpricing · input_price

Output price14 $/Mpricing · output_price

GPQA Diamond92.4 %metric

AIME 2025100 %metric

SWE-Bench Verified80 %metric

Evidence

Citations and source overlap

VellumTop models per benchmark from page-level structured data (JSON-LD).Retrieved 2026-05-20 LLM StatsPublic full LLM leaderboard advanced explorer snapshot, including ranking, benchmark indexes, pricing, speed, context, license, and model metadata.Retrieved 2026-06-01

FAQ

How should I read this profile?

Treat this as a source-backed model dossier, not an EvalKit-run verification. The public values are replicated from linked sources and kept source-scoped.

Is GPT-5.2 verified by EvalKit?

No. EvalKit currently shows 0 verified rows until real run evidence exists.

Why can metrics disagree?

Different sources test different tasks, dates, prompts, and aggregation methods. EvalKit keeps those differences visible instead of merging them into a fake universal score.