Enter opens the highlighted result. Esc closes.

GPT-6 Luna vs Qwen4 Max

Qwen4 Max leads GPT-6 Luna on coding, 74.0 to 54.9 on the Vibe Coding Index, and GPT-6 Luna costs 8.6x less per output token ($0.50 vs $4.275 per million). Updated November 22, 2026; 26 benchmarks have a result for both.

GPT-6 LunaOpenAI
Coding
54.9
Overall
55.9
Output $/M
$0.50
1
Qwen4 MaxQwen
Coding
74.0
Overall
66.6
Output $/M
$4.275
25
vs boards won of 26 shared

GPT-6 Luna or Qwen4 Max: which to pick

Pick Qwen4 Max for coding: it scores 74.0 on the Vibe Coding Index against 54.9 for GPT-6 Luna, though on 4 benchmarks against 5.

Pick GPT-6 Luna for output tokens 8.6x cheaper ($0.50 vs $4.275 per million) and a larger context window (1.05M against 1M tokens).

Shared results
26
GPT-6 Luna only
91
Qwen4 Max only
0
Like-for-like categories
7 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
GPT-6 Luna 54.9 from 5 benchmarks, Qwen4 Max 74.0 from 4 benchmarks.
Widest category
Agentic work, 19.5 points apart on average over 6 shared benchmarks.
Widest single gap
Terminal-Bench 4 (Vals AI): GPT-6 Luna 13.6% against Qwen4 Max 48.5%, 46.5 points apart on a 0 to 100 scale.
Output price
GPT-6 Luna $0.50 against Qwen4 Max $4.275 per million output tokens, 8.6x apart.
Context window
GPT-6 Luna takes 1.05M tokens, Qwen4 Max 1M.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryGPT-6 LunaBarsQwen4 MaxTies
Coding110
Intelligence01
Agentic work06
Reasoning03
Knowledge01
Web design01
General03
All shared boards125
CodingLike-for-likeSix of these boards feed the Vibe Coding Index110
Coding benchmarks for GPT-6 Luna and Qwen4 Max
BenchmarkGPT-6 LunaQwen4 MaxLeader
LiveBench coding70.8%69.8%GPT-6 Luna
LMArena WebDev15781751Qwen4 Max
Terminal-Bench 4 (Vals AI)13.6%48.5%Qwen4 Max
Vibe Code Bench (Vals AI)81.7%86.7%Qwen4 Max
Code migration (Vals AI)42.5%Not measuredOnly GPT-6 Luna measured
ProgramBench (Vals AI)0.5%Not measuredOnly GPT-6 Luna measured
IOI (Vals AI)55.6%Not measuredOnly GPT-6 Luna measured
Terminal-Bench 2.1 (Vals AI)73.0%Not measuredOnly GPT-6 Luna measured
SWE-bench Multimodal (Vals AI)37.5%Not measuredOnly GPT-6 Luna measured
CyberBench (Vals AI)76.3%Not measuredOnly GPT-6 Luna measured
SciCode54.6%Not measuredOnly GPT-6 Luna measured
Terminal-Bench 4.0 (Artificial Analysis run)12.6%48.5%Qwen4 Max
FrontierCode 2.025.2%26.5%Qwen4 Max
FrontierCode42.4%Not measuredOnly GPT-6 Luna measured
MirrorCode26.7%Not measuredOnly GPT-6 Luna measured
ALE-Bench1576.91831.0Qwen4 Max
LMArena Text: coding1434Not measuredOnly GPT-6 Luna measured
LMArena Text: software and IT1422Not measuredOnly GPT-6 Luna measured
LMArena WebDev: React15761785Qwen4 Max
LMArena WebDev: HTML15931752Qwen4 Max
LMArena WebDev: from an image15671675Qwen4 Max
GameBench 255.882.4Qwen4 Max
Terminal-Bench 4.016.4%Not measuredOnly GPT-6 Luna measured
Neither model has a result on the other 29 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
Intelligence benchmarks for GPT-6 Luna and Qwen4 Max
BenchmarkGPT-6 LunaQwen4 MaxLeader
AA Intelligence Index27.942.9Qwen4 Max
Epoch Capabilities Index156.4Not measuredOnly GPT-6 Luna measured
Agentic workLike-for-likePublic boards, shown as published, kept out of the index06
Agentic work benchmarks for GPT-6 Luna and Qwen4 Max
BenchmarkGPT-6 LunaQwen4 MaxLeader
Design Arena: agents building web apps1141Not measuredOnly GPT-6 Luna measured
Design Arena: agents building full-stack apps1140Not measuredOnly GPT-6 Luna measured
Design Arena: agents building mobile apps1115Not measuredOnly GPT-6 Luna measured
Design Arena: agents building Android apps1174Not measuredOnly GPT-6 Luna measured
Design Arena: spreadsheets871Not measuredOnly GPT-6 Luna measured
LiveBench research48.5%84.7%Qwen4 Max
LiveBench browser70.3%80.7%Qwen4 Max
LiveBench long agentic24.7%46.6%Qwen4 Max
SREBench (Vals AI)2.7%Not measuredOnly GPT-6 Luna measured
Finance Agent (Vals AI)49.9%Not measuredOnly GPT-6 Luna measured
Tax Agent Bench (Vals AI)19.9%40.4%Qwen4 Max
Legal Research Bench (Vals AI)30.3%41.4%Qwen4 Max
Harvey's Legal Agent Benchmark (Vals AI)4.2%21.3%Qwen4 Max
Excel Modeling Benchmark (Vals AI)68.5%Not measuredOnly GPT-6 Luna measured
BioMysteryBench (Vals AI)61.5%Not measuredOnly GPT-6 Luna measured
Terminal-Bench-Science (Vals AI)4.3%Not measuredOnly GPT-6 Luna measured
Vals Science Index28.4%Not measuredOnly GPT-6 Luna measured
APEX-Agents44.3%Not measuredOnly GPT-6 Luna measured
SpireClimb0.0%Not measuredOnly GPT-6 Luna measured
LMArena Search1157Not measuredOnly GPT-6 Luna measured
Neither model has a result on the other 24 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index03
Reasoning benchmarks for GPT-6 Luna and Qwen4 Max
BenchmarkGPT-6 LunaQwen4 MaxLeader
LiveBench reasoning81.8%92.2%Qwen4 Max
MysteryMechanism (Vals AI)19.4%Not measuredOnly GPT-6 Luna measured
Humanity's Last Exam (Artificial Analysis run)38.5%Not measuredOnly GPT-6 Luna measured
HLE-Diamond (Artificial Analysis run)22.2%44.9%Qwen4 Max
CritPt v2, hard subset (Artificial Analysis run)42.5%43.0%Qwen4 Max
GPQA Diamond (Epoch AI run)90.5%Not measuredOnly GPT-6 Luna measured
ARC-AGI86.7%Not measuredOnly GPT-6 Luna measured
ARC-AGI-259.3%Not measuredOnly GPT-6 Luna measured
Chess puzzles36.0%Not measuredOnly GPT-6 Luna measured
Mystery game puzzles7.0%Not measuredOnly GPT-6 Luna measured
LMArena Text: hard prompts1408Not measuredOnly GPT-6 Luna measured
LMArena Text: expert1437Not measuredOnly GPT-6 Luna measured
LMArena Text: hard prompts in English1407Not measuredOnly GPT-6 Luna measured
LMArena Vision: diagrams1238Not measuredOnly GPT-6 Luna measured
LMArena Vision: homework1257Not measuredOnly GPT-6 Luna measured
GoBench1119Not measuredOnly GPT-6 Luna measured
Neither model has a result on the other 10 boards here
MathDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
Math benchmarks for GPT-6 Luna and Qwen4 Max
BenchmarkGPT-6 LunaQwen4 MaxLeader
OTIS Mock AIME98.9%Not measuredOnly GPT-6 Luna measured
FrontierMath, tiers 1 to 378.9%Not measuredOnly GPT-6 Luna measured
FrontierMath, tier 456.1%Not measuredOnly GPT-6 Luna measured
ProofBench64.0%Not measuredOnly GPT-6 Luna measured
LMArena Text: math1410Not measuredOnly GPT-6 Luna measured
LMArena Text: mathematical work1424Not measuredOnly GPT-6 Luna measured
Neither model has a result on the other 8 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index01
Knowledge benchmarks for GPT-6 Luna and Qwen4 Max
BenchmarkGPT-6 LunaQwen4 MaxLeader
MedCode (Vals AI)44.7%Not measuredOnly GPT-6 Luna measured
AA-Omniscience-8.327.7Qwen4 Max
SimpleQA Verified41.4%Not measuredOnly GPT-6 Luna measured
LMArena Text: science1387Not measuredOnly GPT-6 Luna measured
LMArena Text: medicine and healthcare1353Not measuredOnly GPT-6 Luna measured
LMArena Text: legal and government1393Not measuredOnly GPT-6 Luna measured
LMArena Text: business and finance1381Not measuredOnly GPT-6 Luna measured
FACTS Benchmark Suite64.8%Not measuredOnly GPT-6 Luna measured
Neither model has a result on the other 13 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index01
Web design benchmarks for GPT-6 Luna and Qwen4 Max
BenchmarkGPT-6 LunaQwen4 MaxLeader
Design Arena: websites1264Not measuredOnly GPT-6 Luna measured
Design Arena: UI components1283Not measuredOnly GPT-6 Luna measured
Design Arena: game dev1236Not measuredOnly GPT-6 Luna measured
Design Arena: data viz1209Not measuredOnly GPT-6 Luna measured
Design Arena: 3D1182Not measuredOnly GPT-6 Luna measured
Design Arena: all code categories1254Not measuredOnly GPT-6 Luna measured
Design Arena: SVG1230Not measuredOnly GPT-6 Luna measured
Design Arena: ASCII art1150Not measuredOnly GPT-6 Luna measured
LiveBench design76.1%79.1%Qwen4 Max
GeneralLike-for-likePublic boards, shown as published, kept out of the index03
General benchmarks for GPT-6 Luna and Qwen4 Max
BenchmarkGPT-6 LunaQwen4 MaxLeader
Vals Index52.3%63.2%Qwen4 Max
LiveBench data analysis73.4%83.4%Qwen4 Max
Public Benefits Bench (Vals AI)57.6%Not measuredOnly GPT-6 Luna measured
SAGE (Vals AI)83.4%Not measuredOnly GPT-6 Luna measured
MedScribe (Vals AI)83.7%Not measuredOnly GPT-6 Luna measured
Long context reasoning (AA-LCR)83.3%Not measuredOnly GPT-6 Luna measured
GDPval-AA14431809Qwen4 Max
DTBench90.1%Not measuredOnly GPT-6 Luna measured
LMCA44.5%Not measuredOnly GPT-6 Luna measured
Furniture assembly44.2%Not measuredOnly GPT-6 Luna measured
LMArena Text1390Not measuredOnly GPT-6 Luna measured
LMArena Text: instruction following1405Not measuredOnly GPT-6 Luna measured
LMArena Text: creative writing1360Not measuredOnly GPT-6 Luna measured
LMArena Vision1210Not measuredOnly GPT-6 Luna measured
LMArena Text: English1392Not measuredOnly GPT-6 Luna measured
LMArena Text: non-English1381Not measuredOnly GPT-6 Luna measured
LMArena Text: Chinese1423Not measuredOnly GPT-6 Luna measured
LMArena Text: German1381Not measuredOnly GPT-6 Luna measured
LMArena Text: Japanese1363Not measuredOnly GPT-6 Luna measured
LMArena Text: Korean1349Not measuredOnly GPT-6 Luna measured
LMArena Text: French1403Not measuredOnly GPT-6 Luna measured
LMArena Text: Spanish1396Not measuredOnly GPT-6 Luna measured
LMArena Text: Russian1390Not measuredOnly GPT-6 Luna measured
LMArena Text: Polish1381Not measuredOnly GPT-6 Luna measured
LMArena Text: multi-turn1395Not measuredOnly GPT-6 Luna measured
LMArena Text: longer queries1404Not measuredOnly GPT-6 Luna measured
LMArena Text: writing and language1383Not measuredOnly GPT-6 Luna measured
LMArena Text: entertainment and media1359Not measuredOnly GPT-6 Luna measured
LMArena Vision: English1208Not measuredOnly GPT-6 Luna measured
LMArena Vision: Chinese1273Not measuredOnly GPT-6 Luna measured
LMArena Vision: reading text1220Not measuredOnly GPT-6 Luna measured
LMArena Vision: humor1167Not measuredOnly GPT-6 Luna measured
LMArena Vision: creative writing1198Not measuredOnly GPT-6 Luna measured
Neither model has a result on the other 16 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-6 Luna and Qwen4 Max
BenchmarkGPT-6 LunaQwen4 MaxLeader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

GPT-6 Luna$0.0003Fits in one request
Qwen4 Max$0.0036Fits in one request

GPT-6 Luna costs less

Repository review

50K input + 3K output tokens

GPT-6 Luna$0.0065Fits in one request
Qwen4 Max$0.084Fits in one request

GPT-6 Luna costs less

Agent loop

200K input + 10K output tokens

GPT-6 Luna$0.025Fits in one request
Qwen4 Max$0.328Fits in one request

GPT-6 Luna costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of GPT-6 Luna and Qwen4 Max
SpecificationOpenAI GPT-6 LunaQwen4 Max
MakerOpenAIQwen
WeightsNo public weights listedNo public weights listed
ReleasedSeptember 22, 2026November 16, 2026
Context window1.05M1M
Input $/Mtok$0.10$1.425
Output $/Mtok$0.50$4.275
Coding index54.974.0
Independent measurements behind it54
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-6 Luna better than Qwen4 Max for coding?

Pick Qwen4 Max for coding: it scores 74.0 on the Vibe Coding Index against 54.9 for GPT-6 Luna, though on 4 benchmarks against 5. Pick GPT-6 Luna for output tokens 8.6x cheaper ($0.50 vs $4.275 per million) and a larger context window (1.05M against 1M tokens).

Which is cheaper, GPT-6 Luna or Qwen4 Max?

GPT-6 Luna costs $0.10 per million input tokens and $0.50 per million output tokens; Qwen4 Max costs $1.425 and $4.275. Cost per solved task is known for neither of them.

Which has the bigger context window, GPT-6 Luna or Qwen4 Max?

GPT-6 Luna, with 1.05M tokens against 1M.

How many benchmarks is the GPT-6 Luna vs Qwen4 Max comparison based on?

26 benchmarks have a result for both models as of November 22, 2026. GPT-6 Luna's coding index rests on 5 independent benchmarks and Qwen4 Max's on 4. Where only one model has a result, the page says so instead of naming a winner.

Compare next

Embed this comparison on your site

Free under CC BY 4.0, while the source link stays visible. Preview the frame · Terms and other embeds

<iframe src="https://llmbenchmark.ai/embed/compare/gpt-6-luna-vs-qwen4-max" title="GPT-6 Luna vs Qwen4 Max, LLMBenchmark.ai" width="100%" height="360" style="border:0;max-width:100%" loading="lazy"></iframe>
<p style="font:13px/1.4 sans-serif;margin:6px 0 0">Source: <a href="https://llmbenchmark.ai/compare/gpt-6-luna-vs-qwen4-max">LLMBenchmark.ai, GPT-6 Luna vs Qwen4 Max</a></p>