BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases9

GPT-5.4 vs Qwen3.7

Full profiles: GPT-5.4 (OpenAI) and Qwen3.7 (Alibaba): every benchmark, price and rank.

Updated · Data from 13 sources

As of 29 Oct 2026, GPT-5.4 is ahead of Qwen3.7 on quality (63.3 vs 60.3 on the BenchLeader Index), but Qwen3.7 is 1.5× cheaper at $3.75 vs $5.63 per 1M tokens blended. Choose GPT-5.4 for agents & tools (64 vs 49), maths (65 vs 58), and coding (65 vs 59); choose Qwen3.7 for fast first answers (14.1 s vs 136.5 s) and instruction following (78 vs 72) and knowledge (61 vs 59), where it also streams ~2× faster.

Key facts (as of 29 Oct 2026)

  • GPT-5.4 scores 63.3 on the BenchLeader Index and Qwen3.7 scores 60.3 (BenchLeader, 29 Oct 2026).
  • GPT-5.4 costs $5.63 per 1M tokens blended (list price, 29 Oct 2026).
  • Qwen3.7 costs $3.75 per 1M tokens blended (list price, 29 Oct 2026).
  • Qwen3.7 streams ~2× faster than GPT-5.4 (Artificial Analysis, 7-day median, 29 Oct 2026).
  • Of the 42 benchmarks both models have, GPT-5.4 leads 31 (BenchLeader, 29 Oct 2026).
GPT-5.4 benchmarks xhigh OpenAI Qwen3.7 benchmarks max AlibabaOpen in the comparison tool

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
CodingGPT-5.465 vs 59 category score; SciCode 56.6% vs 49.5%Medium
Agents & tool useGPT-5.464 vs 49 category score; τ³-Banking (AA) 39.6% vs 11.8%High
MathsGPT-5.465 vs 58 category score; FrontierMath Tiers 1–3 78.6% vs 64.6%Medium
ReasoningEither66 vs 66 category score; GPQA Diamond 93.3% vs 90.9%Either
KnowledgeQwen3.761 vs 59 category score; SimpleQA Verified 55.8% vs 45.1%Medium
Long documentsGPT-5.466 vs 64 category score; AA-LCR v1.1 82.0% vs 79.0%Medium
Following instructionsQwen3.778 vs 72 category score; IFBench 80.5% vs 74.0%Medium
Chat quality (human preference)–Only Qwen3.7 has an Arena rating (56 category score)Not comparable
Lowest costQwen3.7$3.75 vs $5.63 per 1M blended; agentic coding session $0.180 vs $0.210–
Fastest first answerQwen3.714.1 s (Qwen3.7 (max)) vs 136.5 s (GPT-5.4 (xhigh)) 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-5.4 (xhigh) leads on quality: 63.3 vs 60.3.
  • GPT-5.4 (xhigh) is stronger in agents & tools, coding, long context, maths, multimodal.
  • Qwen3.7 (max) is stronger in instruction following, knowledge, reasoning, human preference.
  • Qwen3.7 (max) is 1.5× cheaper ($3.75 vs $5.63 per 1M blended).
  • Qwen3.7 (max) streams 2.1× faster (201 vs 94 tokens per second).

At a glance

SpecGPT-5.4Qwen3.7
DeveloperOpenAIAlibaba
Released5 Mar 202619 May 2026
Context window (tokens)1,050,0001,000,000
Compared atGPT-5.4 (xhigh)Qwen3.7 (max)
Open weightsNoNo
Price in / out / cached input per 1M$2.50 / $15.00 / $0.250$2.50 / $7.50 / $0.500

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-5.4 (xhigh)Qwen3.7 (max)Leader
Reasoning66.166.2level
Coding64.559.1GPT-5.4
Agents & tools63.649.3GPT-5.4
Maths65.158.1GPT-5.4
Knowledge58.661.2Qwen3.7
Instruction following72.378.0Qwen3.7
Human preference–55.6not comparable
Multimodal61.1–not comparable
Long context65.964.3GPT-5.4

Where sources disagree

Across the 42 benchmarks both have, GPT-5.4 leads 31, Qwen3.7 leads 11. Every overall index that rates both puts GPT-5.4 ahead.

Overall indexGPT-5.4 (xhigh)Qwen3.7 (max)Leader
BenchLeader Index63.360.3GPT-5.4
Arena Text1475†1474GPT-5.4

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; suite cost is what Artificial Analysis paid to run its whole Intelligence Index.

MeasureGPT-5.4 (xhigh)Qwen3.7 (max)
Blended price per 1M tokens$5.63$3.75
Input / output per 1M$2.50 / $15.00$2.50 / $7.50
Cached input per 1M$0.250$0.500
Price per task: chat reply$0.0055$0.0033
Price per task: summarize a 30-page report$0.039$0.035
Price per task: code edit$0.037$0.026
Price per task: agentic coding session$0.210$0.180
Price per task: structured extraction$0.0080$0.0065
Output speed94 tok/s201 tok/s
Time to first answer136.5 s14.1 s
Time to a complete answer141.9 s16.6 s

Each model is compared at its best-ranked reasoning-effort setting. GPT-5.4 (xhigh) scores 63.3 and takes 137 s to its first answer token; GPT-5.4 (low) scores 58.9 in 2 s, GPT-5.4 (high) scores 58.8 in 1 s. Qwen3.7 (max) scores 60.3 and takes 14 s to its first answer token. 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

127 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-5.4 (xhigh)Qwen3.7 (max)
BenchLeader IndexBenchLeader63.360.3
Agents & tools scoreBenchLeader63.649.3
Coding scoreBenchLeader64.559.1
Instruction following scoreBenchLeader72.378.0
Knowledge scoreBenchLeader58.661.2
Long context scoreBenchLeader65.964.3
Maths scoreBenchLeader65.158.1
Multimodal scoreBenchLeader61.1–
Reasoning scoreBenchLeader66.166.2
Human preference scoreBenchLeader–55.6
Blended price $/MBenchLeader$5.63$3.75
Output speedBenchLeader94 tok/s201 tok/s
Time to first answerBenchLeader136.5 s14.1 s
Context windowBenchLeader1,050,000 tokens1,000,000 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub93.3%90.9%
Humanity's Last ExamScale AI / CAIS36.2%–
SimpleBenchSimpleBench–70.4%
Arena Hard PromptsArena (LMArena)1487†1495
LiveBench 1.0 ReasoningLiveBench88.1%83.3%
GPQA Diamond (AA)Artificial Analysis92.0%92.3%
Humanity's Last Exam (AA)Artificial Analysis43.7%40.5%
GPQA Diamond (Vals)Vals AI91.7%90.2%
Kagi LLM BenchmarkKagi LLM Benchmark63.8%†–
ARC-AGI-1ARC Prize93.7%–
ARC-AGI-2ARC Prize74.0%–
ARC-AGI-3ARC Prize0.2%†–
CritPtArtificial Analysis23.4%13.4%
Chess PuzzlesEpoch AI Benchmarking Hub48.0%26.0%
Mystery Game PuzzlesEpoch AI Benchmarking Hub37.0%32.0%
CL-benchEpoch AI Benchmarking Hub27.9%–
CL-bench LifeEpoch AI Benchmarking Hub21.7%–
LMCAEpoch AI Benchmarking Hub52.0%44.0%
DTBenchEpoch AI Benchmarking Hub94.4%92.3%
ForecastBenchForecastBench63.1%†–
Coding
SWE-bench Verified (Epoch)Epoch AI Benchmarking Hub76.9%†77.3%
SciCodeSciCode56.6%49.5%
WeirdMLWeirdML77.7%–
GSO-BenchGSO-Bench31.4%–
Arena CodingArena (LMArena)1511†1523
Show all 127 benchmarks
MetricPublisherGPT-5.4 (xhigh)Qwen3.7 (max)
Coding
Arena WebDevArena (LMArena)1401†1515
LiveBench 1.0 CodingLiveBench77.5%74.2%
SciCode (AA)Artificial Analysis–49.5%
LiveCodeBenchVals AI84.1%87.1%
SWE-bench (Vals)Vals AI78.2%68.8%
SWE-Bench ProScale AI SEAL59.1%–
Code MigrationVals AI35.0%13.1%
ProgramBenchVals AI0.0%–
Vibe Code Bench v1.1Vals AI67.4%47.7%
SWE-Bench Pro (private)Scale AI SEAL59.1%–
SWE Atlas: Codebase QnAScale AI SEAL40.8%–
SWE Atlas: RefactoringScale AI SEAL44.3%–
SWE Atlas: Test WritingScale AI SEAL44.4%–
MirrorCodeEpoch AI Benchmarking Hub15.6%†–
DeepSWE v1.1Epoch AI Benchmarking Hub51.8%–
ALE-BenchEpoch AI Benchmarking Hub1607†1189.4
AlgoTuneAlgoTune1.9†–
Terminal-Bench 4.0 (AA)Artificial Analysis–1.5%
Terminal-Bench 2.1 (AA)Artificial Analysis78.3%74.5%
Startrise FrontendStartrise–63.6%
Agents & tools
Terminal-BenchTerminal-Bench81.8%†–
APEX-AgentsMercor52.4%†–
Arena AgentArena (LMArena)–-6.6
LiveBench Agentic CodingLiveBench53.8%43.6%
Terminal-Bench HardArtificial Analysis57.6%50.8%
τ²-Bench Telecom (AA)Artificial Analysis87.1%94.7%
Terminal-Bench 2.1 (Vals)Vals AI–61.0%
MCP AtlasScale AI SEAL70.6%–
HiL-BenchScale AI SEAL9.7%†–
τ³-bench Bankingτ²-bench39.4%–
τ³-Banking (AA)Artificial Analysis39.6%11.8%
ITBench SRE (AA)Artificial Analysis–42.5%
Analyst Agent (AA)Artificial Analysis–18.8%
APEX-Agents (AA)Artificial Analysis33.3%–
Finance Agent v2Vals AI–47.8%
Harvey's Legal Agent Benchmark v1Vals AI0.0%1.7%
Legal Research BenchVals AI–25.5%
SkillsBenchVals AI51.7%–
Tax Agent BenchVals AI–15.8%
Terminal-Bench 2.0 (Vals)Vals AI58.4%†59.2%
DeepResearch BenchDeepResearch Bench35.1%†–
PostTrainBenchEpoch AI Benchmarking Hub19.0%†–
METR Time HorizonsMETR74.3%–
GBAEvalEpoch AI Benchmarking Hub45.1%†0.4%
Vending-Bench 2Andon Labs6144.2†–
AutomationBenchArtificial Analysis–23.0%
GDP.pdf (AA)Artificial Analysis–8.6%
EnterpriseOps-GymArtificial Analysis–45.0%
AA-Briefcase v1.1Artificial Analysis–917
DeepSearchQAGoogle63.7%†–
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub78.6%64.6%
FrontierMath Tier 4Epoch AI Benchmarking Hub49.0%34.1%
OTIS Mock AIMEEpoch AI Benchmarking Hub95.3%95.6%
ProofBenchVals AI56.0%26.0%
LiveBench MathematicsLiveBench94.2%85.3%
AIME (Vals)Vals AI96.7%–
AIME 2026MathArena99.2%–
HMMT February 2026MathArena97.7%–
USAMO 2026MathArena95.2%–
MathArena ApexMathArena54.2%–
Arena MathsArena (LMArena)1460†1482
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub45.1%55.8%
LiveBench 1.0 Data AnalysisLiveBench79.3%71.8%
MMLU-ProVals AI87.5%89.3%
LegalBenchVals AI86.0%84.9%
CorpFinVals AI65.3%63.7%
TaxEvalVals AI74.0%75.3%
MedQAVals AI96.1%–
PRBench FinanceScale AI SEAL45.6%†–
PRBench LegalScale AI SEAL44.4%†–
MultiNRCScale AI SEAL58.3%†–
CaseLaw v2Vals AI63.8%–
Excel Modeling BenchmarkVals AI–57.0%
MedCodeVals AI41.3%38.8%
MedScribeVals AI77.5%79.4%
DrugDiscoveryBenchScale AI SEAL–29.3%†
EBR-benchEpoch AI Benchmarking Hub25.4%9.5%
Instruction following
LiveBench LanguageLiveBench82.6%79.7%
LiveBench Instruction FollowingLiveBench70.2%74.0%
IFBenchArtificial Analysis74.0%80.5%
Arena Instruction FollowingArena (LMArena)1461†1465
Human preference
Arena TextArena (LMArena)1465†1474
EQ-Bench 4EQ-Bench1258†1100
Arena Creative WritingArena (LMArena)1434†1437
Arena Multi-turnArena (LMArena)1480†1482
Arena Longer QueriesArena (LMArena)1477†1490
Multimodal
Arena VisionArena (LMArena)1292†–
MMMU-ProArtificial Analysis78.4%–
VISTAScale AI SEAL50.9%†–
MMMU-Pro (Vals)Vals AI87.5%–
MortgageTaxVals AI68.3%–
VTBScale AI SEAL29.2%†–
Blueprint-Bench 2Epoch AI Benchmarking Hub27.1%†–
Long context
AA-LCR v1.1Artificial Analysis82.0%79.0%
Arena DocumentArena (LMArena)1470†–
MLCRArtificial Analysis–14.4%
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub157.2†154.1
LiveBench 1.0LiveBench78.0%73.1%
AA Intelligence Index v4.3.2Artificial Analysis39.029.5
AA Progress Index v1Artificial Analysis311.6293.8
Vals RSI IndexVals AI15.5%–
Safety & honesty
MASKScale AI SEAL89.7%†–

Data as of 2026-10-29. Best configuration of each model; every score links to its source on the model pages.

What changed

  • Qwen3.7 (max): BenchLeader Index 60.8 → 60.3.
  • GPT-5.4 (xhigh): 1 new result (AA Progress Index v1).
  • Qwen3.7 (max): 1 new result (AA Progress Index v1).
  • Qwen3.7 (max): cached input price moved from $0.33/M to $0.50/M (52% up).
  • Qwen3.7 (max): cost per task moved from $1.15/M to $0.93/M (19% down).
  • Qwen3.7 (max): 1 new result (AA-Briefcase v1.1); 1 score revision; cached input price moved from $0.50/M to $0.33/M (34% down).
  • Qwen3.7 (max): cached input price moved from $0.33/M to $0.50/M (52% up).
  • GPT-5.4 (xhigh): 2 new results (Harvey's Legal Agent Benchmark v1, Vals RSI Index).
  • Qwen3.7 (max): 1 new result (Harvey's Legal Agent Benchmark v1); cached input price moved from $0.50/M to $0.33/M (34% down).
  • Qwen3.7 (max): cached input price moved from $0.33/M to $0.50/M (52% up).

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

GPT-5.4 vs Qwen3.7: questions

Is GPT-5.4 better than Qwen3.7?
GPT-5.4 (xhigh) leads on quality: 63.3 vs 60.3. The BenchLeader Index combines every independent quality benchmark; GPT-5.4 (xhigh) is ahead overall as of 29 Oct 2026, but check the category scores for your use.
GPT-5.4 vs Qwen3.7: which is cheaper, and will that change?
Qwen3.7 is cheaper today: $3.75 against $5.63 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
Is GPT-5.4 or Qwen3.7 better for coding?
GPT-5.4 scores higher in coding (65 vs 59 on the category index, where 50 is average). On LiveCodeBench, the top shared coding benchmark, it is 84.1% for GPT-5.4 against 87.1% for Qwen3.7.
Which is faster, GPT-5.4 or Qwen3.7?
Qwen3.7 streams ~2× faster (201 vs 94 output tokens per second, Artificial Analysis 7-day median).
Is GPT-5.4 or Qwen3.7 better for agents and tool use?
GPT-5.4 scores higher in agentic tasks (64 vs 49 on the category index, where 50 is average). On τ³-Banking (AA), the top shared agents & tools benchmark, it is 39.6% for GPT-5.4 against 11.8% for Qwen3.7.
More questions (4)
Which is better for long documents?
GPT-5.4 scores higher on long-context benchmarks (66 vs 64). Both take about 1.1M tokens: GPT-5.4 1,050,000, Qwen3.7 1,000,000.
Which has lower latency?
Qwen3.7 answers first: 14.1 s against 136.5 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-5.4 vs Qwen3.7?
Across the 42 benchmarks both have, GPT-5.4 leads 31, Qwen3.7 leads 11. Every overall index that rates both puts GPT-5.4 ahead.
When was Qwen3.7 released?
Qwen3.7 was released on 19 May 2026.