BenchLeader

GPT-5 vs Grok 4.3

Full profiles: GPT-5 (OpenAI) and Grok 4.3 (SpaceXAI): every benchmark, price and rank.

Updated · Data from 10 sources

As of 16 Nov 2026, GPT-5 and Grok 4.3 are neck and neck on quality (59.0 vs 56.5 on the BenchLeader Index, inside the ±8.6 margin), and Grok 4.3 is 2.2× cheaper at $1.56 vs $3.44 per 1M tokens blended. Choose GPT-5 for maths (65 vs 46), agents & tools (58 vs 45), and long context (70 vs 61); choose Grok 4.3 for reasoning (63 vs 48), instruction following (79 vs 69), and multimodal (60 vs 57).

Key facts (as of 16 Nov 2026)

  • GPT-5 scores 59.0 on the BenchLeader Index and Grok 4.3 scores 56.5 (BenchLeader, 16 Nov 2026).
  • GPT-5 costs $3.44 per 1M tokens blended (list price, 16 Nov 2026).
  • Grok 4.3 costs $1.56 per 1M tokens blended (list price, 16 Nov 2026).
  • Of the 12 benchmarks both models have, Grok 4.3 leads 9 (BenchLeader, 16 Nov 2026).

Which to pick for…

The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.

Use casePickWhyConfidence
CodingEither57 vs 58 category scoreEither
Agents & tool useGPT-558 vs 45 category score; Terminal-Bench Hard 37.9% vs 37.9%High
MathsGPT-565 vs 46 category score; OTIS Mock AIME 87.2% vs 93.3%High
ReasoningGrok 4.363 vs 48 category score; GPQA Diamond 88.8% vs 85.3%High
Long documentsGPT-570 vs 61 category score; AA-LCR v1.1 76.0% vs 73.0%High
Images and other mediaGrok 4.360 vs 57 category score; MMMU-Pro 78.1% vs 74.3%Medium
Following instructionsGrok 4.379 vs 69 category score; IFBench 81.3% vs 70.6%High
Lowest costGrok 4.3$1.56 vs $3.44 per 1M blended; agentic coding session $0.085 vs $0.115–
Fastest first answerEither22.9 s (GPT-5 (medium)) vs 17.7 s (Grok 4.3 (high)) to the first answer token–
High volume on a budgetNeither flagshipA smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs–

Key differences

  • GPT-5 (medium) and Grok 4.3 (high) are neck and neck on quality (59.0 vs 56.5, within the ±8.6 margin).
  • GPT-5 (medium) is stronger in agents & tools, long context, maths.
  • Grok 4.3 (high) is stronger in coding, instruction following, multimodal, reasoning.
  • Grok 4.3 (high) is 2.2× cheaper ($1.56 vs $3.44 per 1M blended).

At a glance

SpecGPT-5Grok 4.3
DeveloperOpenAISpaceXAI
Released7 Aug 202517 Apr 2026
Context window (tokens)400,0001,000,000
Compared atGPT-5 (medium)Grok 4.3 (high)
Open weightsNoNo
Price in / out / cached input per 1M$1.25 / $10.00 / $0.125$1.25 / $2.50 / $0.200

Category by category

Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.

CategoryGPT-5 (medium)Grok 4.3 (high)Leader
Reasoning47.963.3Grok 4.3
Coding57.457.8level
Agents & tools58.144.8GPT-5
Maths65.446.3GPT-5
Knowledge–50.3not comparable
Instruction following69.478.8Grok 4.3
Multimodal56.660.4Grok 4.3
Long context70.461.0GPT-5

On one chart

Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.

2550 avg75ReasoningCodingAgents & toolsMathsInstructionfollowingLong contextMultimodalGPT-5, Reasoning: 47.9GPT-5, Coding: 57.4GPT-5, Agents & tools: 58.1GPT-5, Maths: 65.4GPT-5, Instruction following: 69.4GPT-5, Long context: 70.4GPT-5, Multimodal: 56.6Grok 4.3, Reasoning: 63.3Grok 4.3, Coding: 57.8Grok 4.3, Agents & tools: 44.8Grok 4.3, Maths: 46.3Grok 4.3, Instruction following: 78.8Grok 4.3, Long context: 61.0Grok 4.3, Multimodal: 60.4

Not drawn, measured for fewer than two of these models: Knowledge, Human preference.

Model, click to add or removeIndex

Where sources disagree

Across the 12 benchmarks both have, GPT-5 leads 2, Grok 4.3 leads 9 and 1 is a tie. Every overall index that rates both puts Grok 4.3 ahead.

Overall indexGPT-5 (medium)Grok 4.3 (high)Leader
BenchLeader Index59.056.5neck and neck
Arena Text1435†1442Grok 4.3

Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).

Price, speed and latency

List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.

MeasureGPT-5 (medium)Grok 4.3 (high)
Blended price per 1M tokens$3.44$1.56
Input / output per 1M$1.25 / $10.00$1.25 / $2.50
Cached input per 1M$0.125$0.200
Price per task: chat reply$0.0035$0.0013
Price per task: summarize a 30-page report$0.021$0.017
Price per task: code edit$0.022$0.011
Price per task: agentic coding session$0.115$0.085
Price per task: structured extraction$0.0045$0.0030
Agentic coding session, input mostly cached$0.064$0.038
Output speed82 tok/s85 tok/s
Time to first answer22.9 s17.7 s
Time to a complete answer, last measured 13 Nov 202666.2 s46.5 s

Each model is compared at its best-ranked reasoning-effort setting. GPT-5 (medium) scores 59.0 and takes 23 s to its first answer token; GPT-5 (high) scores 58.2 in 49 s, GPT-5 scores 56.6 in 5 s. Grok 4.3 (high) scores 56.5 and takes 18 s to its first answer token; Grok 4.3 scores 54.5 in 1 s, Grok 4.3 (no reasoning) scores 46.3 in 1 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.

Every benchmark

147 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.

MetricPublisherGPT-5 (medium)Grok 4.3 (high)
BenchLeader IndexBenchLeader59.056.5
Agents & tools scoreBenchLeader58.144.8
Coding scoreBenchLeader57.457.8
Instruction following scoreBenchLeader69.478.8
Long context scoreBenchLeader70.461.0
Maths scoreBenchLeader65.446.3
Multimodal scoreBenchLeader56.660.4
Reasoning scoreBenchLeader47.963.3
Knowledge scoreBenchLeader–50.3
Blended price $/MBenchLeader$3.44$1.56
Output speedBenchLeader82 tok/s85 tok/s
Time to first answerBenchLeader22.9 s17.7 s
Context windowBenchLeader400,000 tokens1,000,000 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub85.3%88.8%
Humanity's Last ExamScale AI / CAIS25.3%†–
SimpleBenchSimpleBench56.7%†–
Arena Hard PromptsArena (LMArena)1449†1456†
LiveBench 1.0 ReasoningLiveBench–70.8%†
GPQA Diamond (AA)Artificial Analysis84.2%90.1%
Humanity's Last Exam (AA)Artificial Analysis25.4%37.2%
GPQA Diamond (Vals)Vals AI85.6%†91.4%†
Kagi LLM BenchmarkKagi LLM Benchmark72.7%†–
GPQA Diamond (HELM)HELM Capabilities79.1%†–
ARC-AGI-1ARC Prize56.2%–
ARC-AGI-2ARC Prize7.5%–
CritPtArtificial Analysis0.0%8.0%
CritPt v2 hard subset (AA)Artificial Analysis11.5%†–
Arena ExpertArena (LMArena)1445†1454†
Chess PuzzlesEpoch AI Benchmarking Hub32.0%18.0%
Mystery Game PuzzlesEpoch AI Benchmarking Hub15.0%–
LMCAEpoch AI Benchmarking Hub40.0%†38.3%
DTBenchEpoch AI Benchmarking Hub90.7%†90.7%
ForecastBenchForecastBench65.0%†63.2%†
Coding
SWE-bench Verified (Epoch)Epoch AI Benchmarking Hub71.5%–
SciCodeSciCode42.9%†48.3%
WeirdMLWeirdML39.8%†49.9%†
GSO-BenchGSO-Bench6.9%†–
Arena CodingArena (LMArena)1462†1487†
Show all 147 benchmarks
MetricPublisherGPT-5 (medium)Grok 4.3 (high)
Coding
Arena WebDevArena (LMArena)14171357†
LiveBench 1.0 CodingLiveBench–69.9%†
SciCode (AA)Artificial Analysis–48.3%
LiveCodeBenchVals AI85.9%†84.5%
SWE-bench (Vals)Vals AI69.0%†71.4%
SWE-Bench ProScale AI SEAL41.8%†–
Aider PolyglotAider polyglot leaderboard86.7%–
SWE-bench Verified (bash only)SWE-bench65.0%–
SWE-bench Verified (any scaffold)SWE-bench74.4%†–
Code MigrationVals AI–6.8%
ProgramBenchVals AI–0.0%
Vibe Code Bench v1.1Vals AI20.1%†19.4%
SWE-Bench Pro (private)Scale AI SEAL14.9%†–
MirrorCodeEpoch AI Benchmarking Hub4.4%†–
ALE-BenchEpoch AI Benchmarking Hub1162.5†944.2†
AlgoTuneAlgoTune1.7†–
Terminal-Bench 4.0 (AA)Artificial Analysis–0.0%
Terminal-Bench 2.1 (AA)Artificial Analysis35.2%†39.7%
GameBench 2 by reasoning settingDuelLab GameBench 238.0–
Design Arena WebsiteDesign Arena1174†1186†
LiveCodeBench (AA)Artificial Analysis70.3%–
Agents & tools
Terminal-BenchTerminal-Bench49.6%–
GDPvalOpenAI34.8%–
Remote Labor IndexScale AI / CAIS1.7%†–
LiveBench Agentic CodingLiveBench–18.5%†
Terminal-Bench HardArtificial Analysis37.9%37.9%
τ²-Bench Telecom (AA)Artificial Analysis86.5%97.7%
Terminal-Bench 2.1 (Vals)Vals AI–42.0%
GDPval-AA v2.1Artificial Analysis925†945
τ³-Banking (AA)Artificial Analysis22.1%†12.4%
ITBench SRE (AA)Artificial Analysis–32.7%
Analyst Agent (AA)Artificial Analysis–8.8%
APEX-Agents (AA)Artificial Analysis–17.0%
Finance Agent v2Vals AI–37.7%
Harvey's Legal Agent Benchmark v1Vals AI–0.4%
Legal Research BenchVals AI–15.4%
Poker AgentVals AI1103.2%†–
SkillsBenchVals AI–40.6%
Tax Agent BenchVals AI–6.2%
Terminal-Bench 2.0 (Vals)Vals AI37.1%†43.5%†
Arena SearchArena (LMArena)1189†1194†
BALROGBALROG32.8%†–
DeepResearch BenchDeepResearch Bench48.6%–
METR Time HorizonsMETR69.6%–
Vending-Bench 2Andon Labs–35.3†
AutomationBenchArtificial Analysis–1.4%
GDP.pdf (AA)Artificial Analysis–5.8%
EnterpriseOps-GymArtificial Analysis–26.3%
AA-Briefcase v1.1Artificial Analysis–763
Design Arena AgenticDesign Arena–1042†
BrowseComp-PlusBrowseComp-Plus57.6%†–
BrowseCompOpenAI20.1%†–
DeepSearchQAGoogle59.4%†–
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub55.4%†42.8%
FrontierMath Tier 4Epoch AI Benchmarking Hub21.9%†14.6%
OTIS Mock AIMEEpoch AI Benchmarking Hub87.2%93.3%
MATH Level 5Epoch AI Benchmarking Hub97.9%–
ProofBenchVals AI18.0%†11.0%
LiveBench MathematicsLiveBench–84.3%†
AIME (Vals)Vals AI93.4%†–
MGSMVals AI92.8%†–
IMO 2025MathArena38.1%†–
MathArena ApexMathArena1.0%†–
Omni-MATH (HELM)HELM Capabilities64.7%†–
MATH 500Vals AI96.0%†–
Arena MathsArena (LMArena)1411†1419†
AIME 2025 (AA)Artificial Analysis91.7%–
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub50.1%†33.2%
LiveBench 1.0 Data AnalysisLiveBench–55.8%†
MMLU-ProVals AI86.5%†85.8%
LegalBenchVals AI86.0%†84.5%†
CorpFinVals AI61.1%†68.5%†
TaxEvalVals AI73.4%†70.8%†
MedQAVals AI96.3%†–
PRBench FinanceScale AI SEAL51.3%†–
PRBench LegalScale AI SEAL49.0%†–
MultiNRCScale AI SEAL52.1%†–
MMLU-Pro (HELM)HELM Capabilities86.3%†–
CaseLaw v2Vals AI66.5%†79.3%†
Excel Modeling BenchmarkVals AI–18.4%
MedCodeVals AI49.6%†38.1%†
MedScribeVals AI83.7%†74.4%†
Public Benefits Bench v1.1Vals AI–51.7%
EBR-benchEpoch AI Benchmarking Hub12.7%†–
GDP.pdfEpoch AI Benchmarking Hub–8.0%
Global-MMLU-Lite (AA)Artificial Analysis90.4%–
Instruction following
LiveBench LanguageLiveBench–73.6%†
LiveBench Instruction FollowingLiveBench–62.8%†
IFBenchArtificial Analysis70.6%81.3%
MultiChallengeScale AI SEAL63.2%†–
TutorBenchScale AI SEAL55.3%†–
IFEval (HELM)HELM Capabilities87.5%†–
WildBench (HELM)HELM Capabilities85.7%†–
Arena Instruction FollowingArena (LMArena)1415†1415†
Human preference
Arena TextArena (LMArena)1427†1442†
EQ-Bench 4EQ-Bench–1065†
EQ-Bench Creative Writing v3EQ-Bench1626†–
Arena Creative WritingArena (LMArena)1389†1422†
Arena Multi-turnArena (LMArena)1449†1448†
Arena Longer QueriesArena (LMArena)1435†1443†
Lech Mazur Creative WritingLech Mazur86.0%–
Multimodal
Arena VisionArena (LMArena)1232†1229†
MMMU-ProArtificial Analysis74.3%78.1%
VISTAScale AI SEAL49.7%†–
MMMU-Pro (Vals)Vals AI81.5%†83.1%†
MortgageTaxVals AI65.5%†48.3%†
VTBScale AI SEAL17.0%†–
GeoBenchGeoBench81.0%56.0%†
VPCTEpoch AI Benchmarking Hub63.2%–
Blueprint-Bench 2Epoch AI Benchmarking Hub–0.0%†
Long context
Fiction.LiveBench 120kFiction.live96.9%–
AA-LCR v1.1Artificial Analysis76.0%73.0%
MLCRArtificial Analysis–8.3%
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub150†149.9†
LiveBench 1.0LiveBench–62.3%†
AA Intelligence Index v4.3.2Artificial Analysis22.924.9
AA Progress Index v1Artificial Analysis264.4271.2
HELM Capabilities meanHELM Capabilities80.7%†–
Vals Multimodal IndexVals AI–43.3%
Safety & honesty
FORTRESSScale AI SEAL17.0%†–
MASKScale AI SEAL79.3%†–
Metadata
AA Openness IndexArtificial Analysis5.6–

Data as of 2026-11-16. Best configuration of each model; every score links to its source on the model pages.

DuelLab GameBench 2, release gb2-20261111-7b598bb13de4 (https://benchmarks.duellab.org/releases/gb2-20261111-7b598bb13de4/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.

What changed

  • GPT-5 (medium): BenchLeader Index 58.2 → 59.0.
  • GPT-5 (medium): 1 new result (Aider Polyglot).
  • GPT-5 (medium): 2 new results (Global-MMLU-Lite (AA), AIME 2025 (AA)); 1 score revision.
  • Grok 4.3 (high): 2 new results (GDPval-AA v2.1, AA-Briefcase v1.1).
  • GPT-5 (medium): 1 score revision.
  • GPT-5 (medium): 1 score revision.
  • GPT-5 (medium): 1 score revision.
  • Grok 4.3 (high): BenchLeader Index 57.7 → 56.6.
  • GPT-5 (medium): 1 new result (AA Progress Index v1).
  • Grok 4.3 (high): 1 new result (AA Progress Index v1).

Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs

GPT-5 vs Grok 4.3: questions

Is GPT-5 better than Grok 4.3?
GPT-5 (medium) and Grok 4.3 (high) are neck and neck on quality (59.0 vs 56.5, within the ±8.6 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 16 Nov 2026, so check the category scores for your use.
GPT-5 vs Grok 4.3: which is cheaper, and will that change?
Grok 4.3 is cheaper today: $1.56 against $3.44 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
Is GPT-5 or Grok 4.3 better for coding?
They are level in coding (57 vs 58 on the category index).
Which is faster, GPT-5 or Grok 4.3?
Grok 4.3 streams faster (85 vs 82 output tokens per second, Artificial Analysis 7-day median).
Is GPT-5 or Grok 4.3 better for agents and tool use?
GPT-5 scores higher in agentic tasks (58 vs 45 on the category index, where 50 is average). On Terminal-Bench Hard, the top shared agents & tools benchmark, it is 37.9% for GPT-5 against 37.9% for Grok 4.3.
More questions (4)
Which is better for long documents?
GPT-5 scores higher on long-context benchmarks (70 vs 61). Grok 4.3 accepts more context: 1,000,000 against 400,000 tokens for GPT-5.
Which has lower latency?
Grok 4.3 answers first: 17.7 s against 22.9 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
Do independent benchmarks agree on GPT-5 vs Grok 4.3?
Across the 12 benchmarks both have, GPT-5 leads 2, Grok 4.3 leads 9 and 1 is a tie. Every overall index that rates both puts Grok 4.3 ahead.
When was Grok 4.3 released?
Grok 4.3 was released on 17 Apr 2026.