Enter opens the highlighted result. Esc closes.

Claude Fable 5.2 vs Qwen4 Max

Claude Fable 5.2 leads Qwen4 Max on coding, 87.5 to 74.0 on the Vibe Coding Index, and Qwen4 Max costs 9.4x less per output token ($4.275 vs $40.00 per million). Updated November 22, 2026; 25 benchmarks have a result for both.

Claude Fable 5.2Anthropic
Coding
87.5
Overall
89.5
Output $/M
$40.00
21
Qwen4 MaxQwen
Coding
74.0
Overall
66.6
Output $/M
$4.275
4
vs boards won of 25 shared

Claude Fable 5.2 or Qwen4 Max: which to pick

Pick Claude Fable 5.2 for coding: it scores 87.5 on the Vibe Coding Index against 74.0 for Qwen4 Max.

Pick Qwen4 Max for output tokens 9.4x cheaper ($4.275 vs $40.00 per million) and work measured by Tax Agent Bench (Vals AI) or Harvey's Legal Agent Benchmark (Vals AI), where it leads.

Shared results
25
Claude Fable 5.2 only
75
Qwen4 Max only
1
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
Claude Fable 5.2 87.5 from 4 benchmarks, Qwen4 Max 74.0 from 4 benchmarks.
Widest category
Reasoning, 24.1 points apart on average over 3 shared benchmarks.
Widest single gap
LiveBench long agentic: Claude Fable 5.2 82.9% against Qwen4 Max 46.6%, 36.3 points apart on a 0 to 100 scale.
Output price
Claude Fable 5.2 $40.00 against Qwen4 Max $4.275 per million output tokens, 9.4x apart.
Context window
Both take 1M tokens.

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
CategoryClaude Fable 5.2BarsQwen4 MaxTies
Coding100
Intelligence10
Agentic work42
Reasoning21
Knowledge10
Web design10
General21
All shared boards214
CodingLike-for-likeSix of these boards feed the Vibe Coding Index100
Coding benchmarks for Claude Fable 5.2 and Qwen4 Max
BenchmarkClaude Fable 5.2Qwen4 MaxLeader
LiveBench coding79.2%69.8%Claude Fable 5.2
LMArena WebDev18241751Claude Fable 5.2
Terminal-Bench 4 (Vals AI)65.2%48.5%Claude Fable 5.2
Vibe Code Bench (Vals AI)93.8%86.7%Claude Fable 5.2
ProgramBench (Vals AI)19.0%Not measuredOnly Claude Fable 5.2 measured
SWE-bench Multimodal (Vals AI)62.3%Not measuredOnly Claude Fable 5.2 measured
SciCode67.7%Not measuredOnly Claude Fable 5.2 measured
Terminal-Bench 4.0 (Artificial Analysis run)61.6%48.5%Claude Fable 5.2
FrontierCode 2.049.1%26.5%Claude Fable 5.2
FrontierSWE69.9%Not measuredOnly Claude Fable 5.2 measured
ALE-Bench2509.81831.0Claude Fable 5.2
LMArena Text: coding1487Not measuredOnly Claude Fable 5.2 measured
LMArena Text: software and IT1484Not measuredOnly Claude Fable 5.2 measured
LMArena WebDev: React18851785Claude Fable 5.2
LMArena WebDev: HTML17811752Claude Fable 5.2
LMArena WebDev: from an imageNot measured1675Only Qwen4 Max measured
GameBench 299.482.4Claude Fable 5.2
CursorBench 4.061.3%Not measuredOnly Claude Fable 5.2 measured
Neither model has a result on the other 34 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
Intelligence benchmarks for Claude Fable 5.2 and Qwen4 Max
BenchmarkClaude Fable 5.2Qwen4 MaxLeader
AA Intelligence Index64.642.9Claude Fable 5.2
Epoch Capabilities Index173.4Not measuredOnly Claude Fable 5.2 measured
Agentic workLike-for-likePublic boards, shown as published, kept out of the index42
Agentic work benchmarks for Claude Fable 5.2 and Qwen4 Max
BenchmarkClaude Fable 5.2Qwen4 MaxLeader
Design Arena: agents building full-stack apps1411Not measuredOnly Claude Fable 5.2 measured
Design Arena: agents building mobile apps1320Not measuredOnly Claude Fable 5.2 measured
Design Arena: spreadsheets1421Not measuredOnly Claude Fable 5.2 measured
LiveBench research94.5%84.7%Claude Fable 5.2
LiveBench browser86.9%80.7%Claude Fable 5.2
LiveBench long agentic82.9%46.6%Claude Fable 5.2
Tax Agent Bench (Vals AI)32.1%40.4%Qwen4 Max
Legal Research Bench (Vals AI)46.1%41.4%Claude Fable 5.2
Harvey's Legal Agent Benchmark (Vals AI)5.0%21.3%Qwen4 Max
BioMysteryBench (Vals AI)83.7%Not measuredOnly Claude Fable 5.2 measured
Terminal-Bench-Science (Vals AI)55.7%Not measuredOnly Claude Fable 5.2 measured
Vals Science Index63.7%Not measuredOnly Claude Fable 5.2 measured
APEX-Agents80.1%Not measuredOnly Claude Fable 5.2 measured
Remote Labor Index41.3%Not measuredOnly Claude Fable 5.2 measured
PostTrainBench44.6%Not measuredOnly Claude Fable 5.2 measured
SpireClimb15.3%Not measuredOnly Claude Fable 5.2 measured
Vending-Bench 2$9,628Not measuredOnly Claude Fable 5.2 measured
Neither model has a result on the other 27 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index21
Reasoning benchmarks for Claude Fable 5.2 and Qwen4 Max
BenchmarkClaude Fable 5.2Qwen4 MaxLeader
LiveBench reasoning88.8%92.2%Qwen4 Max
MysteryMechanism (Vals AI)51.8%Not measuredOnly Claude Fable 5.2 measured
Humanity's Last Exam (Artificial Analysis run)74.3%Not measuredOnly Claude Fable 5.2 measured
HLE-Diamond (Artificial Analysis run)80.2%44.9%Claude Fable 5.2
CritPt v2, hard subset (Artificial Analysis run)76.5%43.0%Claude Fable 5.2
GPQA Diamond (Epoch AI run)93.9%Not measuredOnly Claude Fable 5.2 measured
ARC-AGI98.5%Not measuredOnly Claude Fable 5.2 measured
ARC-AGI-293.8%Not measuredOnly Claude Fable 5.2 measured
SimpleBench94.9%Not measuredOnly Claude Fable 5.2 measured
Chess puzzles79.0%Not measuredOnly Claude Fable 5.2 measured
Mystery game puzzles90.0%Not measuredOnly Claude Fable 5.2 measured
EBR-bench90.5%Not measuredOnly Claude Fable 5.2 measured
LMArena Text: hard prompts1484Not measuredOnly Claude Fable 5.2 measured
LMArena Text: expert1510Not measuredOnly Claude Fable 5.2 measured
LMArena Text: hard prompts in English1483Not measuredOnly Claude Fable 5.2 measured
LMArena Vision: diagrams1322Not measuredOnly Claude Fable 5.2 measured
LMArena Vision: homework1354Not measuredOnly Claude Fable 5.2 measured
Humanity's Last Exam (Scale run)72.4%Not measuredOnly Claude Fable 5.2 measured
GoBench2564Not measuredOnly Claude Fable 5.2 measured
Neither model has a result on the other 7 boards here
MathDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
Math benchmarks for Claude Fable 5.2 and Qwen4 Max
BenchmarkClaude Fable 5.2Qwen4 MaxLeader
OTIS Mock AIME99.7%Not measuredOnly Claude Fable 5.2 measured
FrontierMath, tiers 1 to 395.1%Not measuredOnly Claude Fable 5.2 measured
FrontierMath, tier 4100.0%Not measuredOnly Claude Fable 5.2 measured
ProofBench100.0%Not measuredOnly Claude Fable 5.2 measured
LMArena Text: math1491Not measuredOnly Claude Fable 5.2 measured
LMArena Text: mathematical work1504Not measuredOnly Claude Fable 5.2 measured
MathArena82.1%Not measuredOnly Claude Fable 5.2 measured
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index10
Knowledge benchmarks for Claude Fable 5.2 and Qwen4 Max
BenchmarkClaude Fable 5.2Qwen4 MaxLeader
AA-Omniscience66.827.7Claude Fable 5.2
SimpleQA Verified94.0%Not measuredOnly Claude Fable 5.2 measured
LMArena Text: science1494Not measuredOnly Claude Fable 5.2 measured
LMArena Text: medicine and healthcare1464Not measuredOnly Claude Fable 5.2 measured
LMArena Text: legal and government1472Not measuredOnly Claude Fable 5.2 measured
LMArena Text: business and finance1443Not measuredOnly Claude Fable 5.2 measured
PRBench Finance47.1%Not measuredOnly Claude Fable 5.2 measured
PRBench Legal47.1%Not measuredOnly Claude Fable 5.2 measured
Neither model has a result on the other 13 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index10
Web design benchmarks for Claude Fable 5.2 and Qwen4 Max
BenchmarkClaude Fable 5.2Qwen4 MaxLeader
Design Arena: websites1348Not measuredOnly Claude Fable 5.2 measured
Design Arena: UI components1377Not measuredOnly Claude Fable 5.2 measured
Design Arena: game dev1536Not measuredOnly Claude Fable 5.2 measured
Design Arena: 3D1492Not measuredOnly Claude Fable 5.2 measured
Design Arena: all code categories1392Not measuredOnly Claude Fable 5.2 measured
LiveBench design85.1%79.1%Claude Fable 5.2
Neither model has a result on the other 3 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index21
General benchmarks for Claude Fable 5.2 and Qwen4 Max
BenchmarkClaude Fable 5.2Qwen4 MaxLeader
Vals Index71.3%63.2%Claude Fable 5.2
LiveBench data analysis81.5%83.4%Qwen4 Max
Long context reasoning (AA-LCR)83.3%Not measuredOnly Claude Fable 5.2 measured
GDPval-AA18311809Claude Fable 5.2
DTBench99.2%Not measuredOnly Claude Fable 5.2 measured
LMCA70.6%Not measuredOnly Claude Fable 5.2 measured
Furniture assembly90.0%Not measuredOnly Claude Fable 5.2 measured
LMArena Text1478Not measuredOnly Claude Fable 5.2 measured
LMArena Text: instruction following1490Not measuredOnly Claude Fable 5.2 measured
LMArena Text: creative writing1516Not measuredOnly Claude Fable 5.2 measured
LMArena Vision1313Not measuredOnly Claude Fable 5.2 measured
LMArena Text: English1474Not measuredOnly Claude Fable 5.2 measured
LMArena Text: non-English1474Not measuredOnly Claude Fable 5.2 measured
LMArena Text: Chinese1528Not measuredOnly Claude Fable 5.2 measured
LMArena Text: French1524Not measuredOnly Claude Fable 5.2 measured
LMArena Text: Spanish1446Not measuredOnly Claude Fable 5.2 measured
LMArena Text: Russian1501Not measuredOnly Claude Fable 5.2 measured
LMArena Text: multi-turn1458Not measuredOnly Claude Fable 5.2 measured
LMArena Text: longer queries1499Not measuredOnly Claude Fable 5.2 measured
LMArena Text: writing and language1498Not measuredOnly Claude Fable 5.2 measured
LMArena Text: entertainment and media1486Not measuredOnly Claude Fable 5.2 measured
LMArena Vision: English1315Not measuredOnly Claude Fable 5.2 measured
LMArena Vision: reading text1327Not measuredOnly Claude Fable 5.2 measured
MultiChallenge80.4%Not measuredOnly Claude Fable 5.2 measured
Neither model has a result on the other 25 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Claude Fable 5.2 and Qwen4 Max
BenchmarkClaude Fable 5.2Qwen4 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

Claude Fable 5.2$0.028Fits in one request
Qwen4 Max$0.0036Fits in one request

Qwen4 Max costs less

Repository review

50K input + 3K output tokens

Claude Fable 5.2$0.520Fits in one request
Qwen4 Max$0.084Fits in one request

Qwen4 Max costs less

Agent loop

200K input + 10K output tokens

Claude Fable 5.2$2.00Fits in one request
Qwen4 Max$0.328Fits in one request

Qwen4 Max 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 Claude Fable 5.2 and Qwen4 Max
SpecificationAnthropic Claude Fable 5.2Qwen4 Max
MakerAnthropicQwen
WeightsNo public weights listedNo public weights listed
ReleasedNovember 9, 2026November 16, 2026
Context window1M1M
Input $/Mtok$8.00$1.425
Output $/Mtok$40.00$4.275
Coding index87.574.0
Independent measurements behind it44
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Fable 5.2 better than Qwen4 Max for coding?

Pick Claude Fable 5.2 for coding: it scores 87.5 on the Vibe Coding Index against 74.0 for Qwen4 Max. Pick Qwen4 Max for output tokens 9.4x cheaper ($4.275 vs $40.00 per million) and work measured by Tax Agent Bench (Vals AI) or Harvey's Legal Agent Benchmark (Vals AI), where it leads.

Which is cheaper, Claude Fable 5.2 or Qwen4 Max?

Claude Fable 5.2 costs $8.00 per million input tokens and $40.00 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, Claude Fable 5.2 or Qwen4 Max?

Neither: both take 1M tokens.

How many benchmarks is the Claude Fable 5.2 vs Qwen4 Max comparison based on?

25 benchmarks have a result for both models as of November 22, 2026. Claude Fable 5.2's coding index rests on 4 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/claude-fable-5-2-vs-qwen4-max" title="Claude Fable 5.2 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/claude-fable-5-2-vs-qwen4-max">LLMBenchmark.ai, Claude Fable 5.2 vs Qwen4 Max</a></p>