Enter opens the highlighted result. Esc closes.

Llama 4 Scout vs Qwen3.5 397B A17B

Qwen3.5 397B A17B leads Llama 4 Scout on the Artificial Analysis Intelligence Index, 21.4 to 6.5, and Llama 4 Scout costs 4.7x less per output token ($0.50 vs $2.34 per million). Updated October 22, 2026; 54 benchmarks have a result for both.

Replace Llama 4 Scout Replace Qwen3.5 397B A17B
Llama 4 ScoutMeta Llama
Coding
No data
Overall
24.9
Output $/M
$0.50
0
Qwen3.5 397B A17BQwen
Coding
No data
Overall
57.1
Output $/M
$2.34
54
vs boards won of 54 shared

Llama 4 Scout or Qwen3.5 397B A17B: which to pick

Neither model has a coding score yet, so the two cannot be ranked on coding.

Pick Llama 4 Scout for output tokens 4.7x cheaper ($0.50 vs $2.34 per million) and a larger context window (1.31M against 262K tokens).

Model ALlama 4 ScoutMeta LlamaNot benchmarkedNo coding benchmark$1.276 per solved taskModel BQwen3.5 397B A17BQwenNot benchmarked1 independent benchmarkNo cost per solved task
Shared results
54
Llama 4 Scout only
28
Qwen3.5 397B A17B only
11
Like-for-like categories
8 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
Llama 4 Scout not benchmarked, Qwen3.5 397B A17B not benchmarked.
Widest category
Agentic work, 80.1 points apart on average over 1 shared benchmark.
Widest single gap
OTIS Mock AIME: Llama 4 Scout 7.8% against Qwen3.5 397B A17B 88.9%, 81.1 points apart on a 0 to 100 scale.
Output price
Llama 4 Scout $0.50 against Qwen3.5 397B A17B $2.34 per million output tokens, 4.7x apart.
Context window
Llama 4 Scout takes 1.31M tokens, Qwen3.5 397B A17B 262K.

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
CategoryLlama 4 ScoutBarsQwen3.5 397B A17BTies
Coding04
Intelligence02
Agentic work01
Reasoning08
Math03
Knowledge04
Web design05
General027
All shared boards054
CodingLike-for-likeSix of these boards feed the Vibe Coding Index04
Coding benchmarks for Llama 4 Scout and Qwen3.5 397B A17B
BenchmarkLlama 4 ScoutQwen3.5 397B A17BLeader
SWE-bench Verified9.1%Not measuredOnly Llama 4 Scout measured
LMArena WebDevNot measured1400Only Qwen3.5 397B A17B measured
AA Coding Index8.2Not measuredOnly Llama 4 Scout measured
LiveCodeBench (Vals AI)38.5%Not measuredOnly Llama 4 Scout measured
LiveCodeBench (Artificial Analysis run)29.9%Not measuredOnly Llama 4 Scout measured
SciCode21.3%44.8%Qwen3.5 397B A17B
Terminal-Bench Hard (Artificial Analysis run)1.5%40.9%Qwen3.5 397B A17B
ALE-BenchNot measured621.9Only Qwen3.5 397B A17B measured
LMArena Text: coding12861464Qwen3.5 397B A17B
LMArena Text: software and IT12861463Qwen3.5 397B A17B
LMArena WebDev: ReactNot measured1388Only Qwen3.5 397B A17B measured
LMArena WebDev: HTMLNot measured1410Only Qwen3.5 397B A17B measured
Neither model has a result on the other 27 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for Llama 4 Scout and Qwen3.5 397B A17B
BenchmarkLlama 4 ScoutQwen3.5 397B A17BLeader
AA Intelligence Index6.521.4Qwen3.5 397B A17B
Epoch Capabilities Index129.6146.8Qwen3.5 397B A17B
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
Agentic work benchmarks for Llama 4 Scout and Qwen3.5 397B A17B
BenchmarkLlama 4 ScoutQwen3.5 397B A17BLeader
AA Agentic Index0.5Not measuredOnly Llama 4 Scout measured
τ²-Bench Telecom15.5%95.6%Qwen3.5 397B A17B
APEX-AgentsNot measured24.9%Only Qwen3.5 397B A17B measured
Berkeley Function Calling Leaderboard (BFCL)28.1%Not measuredOnly Llama 4 Scout measured
Neither model has a result on the other 37 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index08
Reasoning benchmarks for Llama 4 Scout and Qwen3.5 397B A17B
BenchmarkLlama 4 ScoutQwen3.5 397B A17BLeader
GPQA Diamond47.0%Not measuredOnly Llama 4 Scout measured
MMMU Pro (Vals AI)58.8%Not measuredOnly Llama 4 Scout measured
GPQA Diamond (Artificial Analysis run)58.7%89.3%Qwen3.5 397B A17B
Humanity's Last Exam (Artificial Analysis run)3.8%29.0%Qwen3.5 397B A17B
GPQA Diamond (Epoch AI run)51.8%86.4%Qwen3.5 397B A17B
ARC-AGI0.5%Not measuredOnly Llama 4 Scout measured
ARC-AGI-20.0%Not measuredOnly Llama 4 Scout measured
Mystery game puzzlesNot measured18.0%Only Qwen3.5 397B A17B measured
LMArena Text: hard prompts12661448Qwen3.5 397B A17B
LMArena Text: expert12351462Qwen3.5 397B A17B
LMArena Text: hard prompts in English12931450Qwen3.5 397B A17B
LMArena Vision: diagrams11251277Qwen3.5 397B A17B
LMArena Vision: homework11481292Qwen3.5 397B A17B
Kagi LLM Benchmark36.9%Not measuredOnly Llama 4 Scout measured
Neither model has a result on the other 8 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index03
Math benchmarks for Llama 4 Scout and Qwen3.5 397B A17B
BenchmarkLlama 4 ScoutQwen3.5 397B A17BLeader
AIME19.0%Not measuredOnly Llama 4 Scout measured
MATH 500 (Vals AI)79.2%Not measuredOnly Llama 4 Scout measured
MGSM (Vals AI)88.0%Not measuredOnly Llama 4 Scout measured
AA Math Index14.0Not measuredOnly Llama 4 Scout measured
AIME 2025 (Artificial Analysis run)14.0%Not measuredOnly Llama 4 Scout measured
OTIS Mock AIME7.8%88.9%Qwen3.5 397B A17B
FrontierMath, tiers 1 to 3Not measured31.2%Only Qwen3.5 397B A17B measured
MATH level 562.3%Not measuredOnly Llama 4 Scout measured
LMArena Text: math12871453Qwen3.5 397B A17B
LMArena Text: mathematical work12861454Qwen3.5 397B A17B
Neither model has a result on the other 3 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index04
Knowledge benchmarks for Llama 4 Scout and Qwen3.5 397B A17B
BenchmarkLlama 4 ScoutQwen3.5 397B A17BLeader
MMLU Pro69.6%Not measuredOnly Llama 4 Scout measured
LegalBench (Vals AI)72.0%Not measuredOnly Llama 4 Scout measured
CorpFin (Vals AI)46.8%Not measuredOnly Llama 4 Scout measured
TaxEval (Vals AI)55.2%Not measuredOnly Llama 4 Scout measured
MortgageTax (Vals AI)57.8%Not measuredOnly Llama 4 Scout measured
MedQA (Vals AI)50.9%Not measuredOnly Llama 4 Scout measured
MedCode (Vals AI)23.3%Not measuredOnly Llama 4 Scout measured
MMLU Pro (Artificial Analysis run)75.2%Not measuredOnly Llama 4 Scout measured
LMArena Text: science12781457Qwen3.5 397B A17B
LMArena Text: medicine and healthcare12751455Qwen3.5 397B A17B
LMArena Text: legal and government12851441Qwen3.5 397B A17B
LMArena Text: business and finance12571435Qwen3.5 397B A17B
Neither model has a result on the other 6 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index05
Web design benchmarks for Llama 4 Scout and Qwen3.5 397B A17B
BenchmarkLlama 4 ScoutQwen3.5 397B A17BLeader
Design Arena: websites7501190Qwen3.5 397B A17B
Design Arena: UI components7651161Qwen3.5 397B A17B
Design Arena: game dev7721140Qwen3.5 397B A17B
Design Arena: data viz8961175Qwen3.5 397B A17B
Design Arena: 3DNot measured1160Only Qwen3.5 397B A17B measured
Design Arena: all code categories7891179Qwen3.5 397B A17B
Design Arena: SVGNot measured1139Only Qwen3.5 397B A17B measured
Neither model has a result on the other 2 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index027
General benchmarks for Llama 4 Scout and Qwen3.5 397B A17B
BenchmarkLlama 4 ScoutQwen3.5 397B A17BLeader
SAGE (Vals AI)34.8%Not measuredOnly Llama 4 Scout measured
MedScribe (Vals AI)50.6%Not measuredOnly Llama 4 Scout measured
IFBench39.5%78.8%Qwen3.5 397B A17B
Long context reasoning (AA-LCR)27.7%77.3%Qwen3.5 397B A17B
DTBench57.9%87.5%Qwen3.5 397B A17B
LMCA12.0%37.9%Qwen3.5 397B A17B
Fiction.LiveBench36.0%Not measuredOnly Llama 4 Scout measured
LMArena Text12791437Qwen3.5 397B A17B
LMArena Text: instruction following12481424Qwen3.5 397B A17B
LMArena Text: creative writing12491402Qwen3.5 397B A17B
LMArena Vision11181262Qwen3.5 397B A17B
LMArena Text: English13041443Qwen3.5 397B A17B
LMArena Text: non-English12521430Qwen3.5 397B A17B
LMArena Text: Chinese12551498Qwen3.5 397B A17B
LMArena Text: German12731445Qwen3.5 397B A17B
LMArena Text: Japanese12061429Qwen3.5 397B A17B
LMArena Text: Korean12071384Qwen3.5 397B A17B
LMArena Text: French12831462Qwen3.5 397B A17B
LMArena Text: Spanish12781441Qwen3.5 397B A17B
LMArena Text: Russian12631430Qwen3.5 397B A17B
LMArena Text: Polish12651447Qwen3.5 397B A17B
LMArena Text: multi-turn12801446Qwen3.5 397B A17B
LMArena Text: longer queries12651441Qwen3.5 397B A17B
LMArena Text: writing and language12691415Qwen3.5 397B A17B
LMArena Text: entertainment and media12421399Qwen3.5 397B A17B
LMArena Vision: English11431268Qwen3.5 397B A17B
LMArena Vision: Chinese11061314Qwen3.5 397B A17B
LMArena Vision: reading text11171271Qwen3.5 397B A17B
LMArena Vision: humor11021262Qwen3.5 397B A17B
LMArena Vision: captioningNot measured1230Only Qwen3.5 397B A17B measured
LMArena Vision: recognizing thingsNot measured1265Only Qwen3.5 397B A17B measured
LMArena Vision: creative writing11261260Qwen3.5 397B A17B
Neither model has a result on the other 14 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Llama 4 Scout and Qwen3.5 397B A17B
BenchmarkLlama 4 ScoutQwen3.5 397B A17BLeader
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

Llama 4 Scout$0.0004Fits in one request
Qwen3.5 397B A17B$0.0016Fits in one request

Llama 4 Scout costs less

Repository review

50K input + 3K output tokens

Llama 4 Scout$0.010Fits in one request
Qwen3.5 397B A17B$0.027Fits in one request

Llama 4 Scout costs less

Agent loop

200K input + 10K output tokens

Llama 4 Scout$0.040Fits in one request
Qwen3.5 397B A17B$0.101Fits in one request

Llama 4 Scout 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 Llama 4 Scout and Qwen3.5 397B A17B
SpecificationMeta Llama 4 ScoutQwen3.5 397B A17B
MakerMeta LlamaQwen
WeightsOpen weightsOpen weights
ReleasedApril 5, 2025February 16, 2026
Context window1.31M262K
Input $/Mtok$0.175$0.39
Output $/Mtok$0.50$2.34
Coding indexNot benchmarkedNot benchmarked
Independent measurements behind it01
Cost per solved task$1.276Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Llama 4 Scout better than Qwen3.5 397B A17B for coding?

Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Llama 4 Scout for output tokens 4.7x cheaper ($0.50 vs $2.34 per million) and a larger context window (1.31M against 262K tokens).

Which is cheaper, Llama 4 Scout or Qwen3.5 397B A17B?

Llama 4 Scout costs $0.175 per million input tokens and $0.50 per million output tokens; Qwen3.5 397B A17B costs $0.39 and $2.34. Cost per solved task is known for Llama 4 Scout only.

Which has the bigger context window, Llama 4 Scout or Qwen3.5 397B A17B?

Llama 4 Scout, with 1.31M tokens against 262K.

How many benchmarks is the Llama 4 Scout vs Qwen3.5 397B A17B comparison based on?

54 benchmarks have a result for both models as of October 22, 2026. Llama 4 Scout's coding index rests on 0 independent benchmarks and Qwen3.5 397B A17B's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next