Model comparison

GPT-5.2 vs Grok 4.20 (Non-Reasoning)

GPT-5.2 is the stronger model overall, scoring 54.1 to 48.6 on the Noometry Index. Grok 4.20 (Non-Reasoning) costs 3.1× less per token, which makes it the better buy when GPT-5.2's lead doesn't matter for your workload.

Last verified October 10, 2026. 43 shared benchmarks.

GPT-5.2 OpenAI

54.1

Rank #34 Confirmed

Grok 4.20 (Non-Reasoning) xAI

48.6

Rank #54 Confirmed

Summary

They share 43 benchmarks with published results for both. GPT-5.2 scores higher in 6 categories and Grok 4.20 (Non-Reasoning) in 4 categories; 9 gaps are clear of the uncertainty.
The widest gap is in multimodal, where GPT-5.2 leads 51.3 to 33.3.
The biggest single-benchmark swing is Chess Puzzles: 49% for GPT-5.2 and 24% for Grok 4.20 (Non-Reasoning).
Grok 4.20 (Non-Reasoning) is cheaper at $1.25 / $2.50 per million input/output tokens, against $1.75 / $14 for GPT-5.2.
Grok 4.20 (Non-Reasoning) accepts more context: 1M tokens versus 400K.

Side by side

GPT-5.2 and Grok 4.20 (Non-Reasoning) specifications
	GPT-5.2	Grok 4.20 (Non-Reasoning)
Provider	OpenAI	xAI
Noometry Index	54.1	48.6
Released	2025-12-11	2026-02-17
Weights	Proprietary	Proprietary
Context window	400K	1M
Max output	128K	30K
Input $ / M tokens	$1.75	$1.25
Output $ / M tokens	$14	$2.50
Results tracked	67	46

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding GPT-5.2 leads

GPT-5.2: 51.6 (#37), Grok 4.20 (Non-Reasoning): 42.1 (#112)

Coding benchmarks
Benchmark	GPT-5.2	Grok 4.20 (Non-Reasoning)
LMArena WebDev	1416	1375
WeirdML	72.2%	52.3%
LMArena Coding	1447	1459
ALE-Bench	1,294	1,150
SWE-bench Verified	73.8%	—
SWE-bench Verified (bash only)	72.8%	—
SWE-bench Multilingual	66.7%	—
GSO	27.4%	—
AlgoTune	2.05	—

Agentic & Tool Use GPT-5.2 leads

GPT-5.2: 40.2 (#24), Grok 4.20 (Non-Reasoning): 34.4 (#46)

Agentic & Tool Use benchmarks
Benchmark	GPT-5.2	Grok 4.20 (Non-Reasoning)
Terminal-Bench	64.9%	57.3%
τ²-bench Banking	32.2%	18%
LMArena Search	1207	1189
Vending-Bench 2	3,591	4,663
Berkeley Function Calling Leaderboard	55.9%	—
GDPval	49.7%	—
Remote Labor Index	2.5%	—
τ²-bench Airline	83%	—
τ²-bench Retail	81.6%	—
τ²-bench Telecom	89.7%	—
DeepResearch Bench	41.1%	—
METR Time Horizons	75.3%	—

Reasoning Grok 4.20 (Non-Reasoning) leads

GPT-5.2: 50.2 (#35), Grok 4.20 (Non-Reasoning): 52.3 (#32)

Reasoning benchmarks
Benchmark	GPT-5.2	Grok 4.20 (Non-Reasoning)
ARC-AGI-2	52.9%	65.1%
Kagi LLM Benchmark	73.3%	75%
NYT Connections (extended)	83.6%	85.4%
ARC-AGI-1	86.2%	89.5%
Chess Puzzles	49%	24%
LMArena Hard Prompts	1445	1451
DTBench	90.9%	90.1%
LMCA	43.9%	38.7%
Epoch Capabilities Index	153.45	151.98
ForecastBench	60.1	61.4
SimpleBench	45.8%	—
EnigmaEval	10.4%	—
Thematic Generalization	—	63.8%
EBR-Bench	23%	—
Mystery Game Puzzles	23%	—

Math GPT-5.2 leads

GPT-5.2: 60.0 (#38), Grok 4.20 (Non-Reasoning): 48.2 (#65)

Math benchmarks
Benchmark	GPT-5.2	Grok 4.20 (Non-Reasoning)
FrontierMath (Tiers 1-3)	67.4%	44.9%
FrontierMath Tier 4	31.7%	17.1%
OTIS Mock AIME 2024-2025	96.1%	92.2%
ProofBench	15%	14%
LMArena Math	1440	1455
MathArena Final-Answer Competitions	72%	—
FrontierMath (Feb 2025 set)	40.7%	—
FrontierMath Tier 4 (v1)	18.8%	—

Knowledge GPT-5.2 leads

GPT-5.2: 59.3 (#32), Grok 4.20 (Non-Reasoning): 52.8 (#60)

Knowledge benchmarks
Benchmark	GPT-5.2	Grok 4.20 (Non-Reasoning)
GPQA Diamond	91.4%	89.3%
SimpleQA Verified	37.1%	30.2%
LMArena Expert	1445	1439
Humanity's Last Exam	27.8%	—
Vectara Hallucination Rate	8.4%	—

Multimodal GPT-5.2 leads

GPT-5.2: 51.3 (#7), Grok 4.20 (Non-Reasoning): 33.3 (#98)

Multimodal benchmarks
Benchmark	GPT-5.2	Grok 4.20 (Non-Reasoning)
LMArena Vision	1268	1263
LMArena Document	1405	1416
VPCT	84%	—
Blueprint-Bench 2	—	0%
Furniture Assembly	38.3%	—

Multilingual Grok 4.20 (Non-Reasoning) leads

GPT-5.2: 53.4 (#67), Grok 4.20 (Non-Reasoning): 54.5 (#40)

Multilingual benchmarks
Benchmark	GPT-5.2	Grok 4.20 (Non-Reasoning)
LMArena Non-English	1425	1441
LMArena Chinese	1460	1481
LMArena French	1455	1476
LMArena German	1448	1465
LMArena Japanese	1420	1449
LMArena Korean	1392	1417
LMArena Russian	1440	1458
LMArena Spanish	1433	1443

Instruction Following Too close to call

GPT-5.2: 74.7 (#89), Grok 4.20 (Non-Reasoning): 74.8 (#83)

Instruction Following benchmarks
Benchmark	GPT-5.2	Grok 4.20 (Non-Reasoning)
LMArena Instruction Following	1417	1420

Long Context Grok 4.20 (Non-Reasoning) leads

GPT-5.2: 44.0 (#78), Grok 4.20 (Non-Reasoning): 45.5 (#34)

Long Context benchmarks
Benchmark	GPT-5.2	Grok 4.20 (Non-Reasoning)
CL-bench	18.2%	22.2%
LMArena Longer Query	1428	1437
CL-bench Life	—	11.9%

Writing & Preference GPT-5.2 leads

GPT-5.2: 66.8 (#32), Grok 4.20 (Non-Reasoning): 65.7 (#44)

Writing & Preference benchmarks
Benchmark	GPT-5.2	Grok 4.20 (Non-Reasoning)
LMArena Text	1439	1451
LMArena Creative Writing	1401	1438
EQ-Bench Creative Writing	1703	1574
LMArena Multi-Turn	1458	1456

Frequently asked questions

Is GPT-5.2 better than Grok 4.20 (Non-Reasoning)?

Which is cheaper, GPT-5.2 or Grok 4.20 (Non-Reasoning)?

Grok 4.20 (Non-Reasoning) is cheaper. It lists at $1.25 per million input tokens and $2.50 per million output tokens; GPT-5.2 lists at $1.75 and $14.

Is GPT-5.2 or Grok 4.20 (Non-Reasoning) better for coding?

GPT-5.2 scores higher on coding benchmarks: 51.6 versus 42.1 in the Noometry coding category.

Which has the bigger context window?

Grok 4.20 (Non-Reasoning) does, with 1M tokens against 400K.

How many benchmarks do GPT-5.2 and Grok 4.20 (Non-Reasoning) share?

43 benchmarks have published results for both models. GPT-5.2 has 67 scored results on Noometry and Grok 4.20 (Non-Reasoning) has 46.