LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Gemma 4 31B vs Qwen3.5 397B A17B

Updated October 21, 2026. 52 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace Gemma 4 31B Replace Qwen3.5 397B A17B
Gemma 4 31BGoogle
Coding
38.3
Overall
14.7
Output $/M
$0.40
28
Qwen3.5 397B A17BQwen
Coding
28.6
Overall
21.4
Output $/M
$2.34
24
vs boards won of 52 shared

Decision reading

Pick Gemma 4 31B for coding: it scores 38.3 on the Vibe Coding Index against 28.6 for Qwen3.5 397B A17B.

Pick Qwen3.5 397B A17B when what matters is LMArena WebDev or Terminal-Bench Hard (Artificial Analysis run), where it leads.

Model AGemma 4 31BGoogle38.3/1002 independent benchmarksNo cost per solved taskModel BQwen3.5 397B A17BQwen28.6/1001 independent benchmarkNo cost per solved task
Shared results
52
Gemma 4 31B only
16
Qwen3.5 397B A17B only
13
Like-for-like categories
7 of 9

Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.

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
Gemma 4 31B 38.3 from 2 benchmarks, Qwen3.5 397B A17B 28.6 from 1 benchmark.
Widest category
Agentic work, 30.1 points apart on average over 1 shared benchmark.
Widest single gap
τ²-Bench Telecom: Gemma 4 31B 65.5% against Qwen3.5 397B A17B 95.6%, 30.1 points apart on a 0 to 100 scale.
Output price
Gemma 4 31B $0.40 against Qwen3.5 397B A17B $2.34 per million output tokens, 5.8x apart.
Context window
Both take 262K 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
CategoryGemma 4 31BBarsQwen3.5 397B A17BTies
Coding35
Intelligence02
Agentic work01
Reasoning44
Math21
Knowledge13
General188
All shared boards2824
CodingLike-for-likeFour of these boards feed the Vibe Coding Index35
Coding benchmarks for Gemma 4 31B and Qwen3.5 397B A17B
BenchmarkGemma 4 31BQwen3.5 397B A17BLeader
LiveBench coding49.0%Not measuredOnly Gemma 4 31B measured
LMArena WebDev13671400Qwen3.5 397B A17B
AA Coding Index43.4Not measuredOnly Gemma 4 31B measured
SciCode45.5%44.8%Gemma 4 31B
Terminal-Bench Hard (Artificial Analysis run)36.4%40.9%Qwen3.5 397B A17B
WeirdML52.3%Not measuredOnly Gemma 4 31B measured
ALE-Bench925.5621.9Gemma 4 31B
LMArena Text: coding14591464Qwen3.5 397B A17B
LMArena Text: software and IT14631463Gemma 4 31B
LMArena WebDev: React13591388Qwen3.5 397B A17B
LMArena WebDev: HTML13801410Qwen3.5 397B A17B
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 Gemma 4 31B and Qwen3.5 397B A17B
BenchmarkGemma 4 31BQwen3.5 397B A17BLeader
AA Intelligence Index14.721.4Qwen3.5 397B A17B
Epoch Capabilities Index143.0146.8Qwen3.5 397B A17B
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
Agentic work benchmarks for Gemma 4 31B and Qwen3.5 397B A17B
BenchmarkGemma 4 31BQwen3.5 397B A17BLeader
AA Agentic Index4.2Not measuredOnly Gemma 4 31B measured
LiveBench research44.4%Not measuredOnly Gemma 4 31B measured
LiveBench browser4.0%Not measuredOnly Gemma 4 31B measured
LiveBench long agentic13.3%Not measuredOnly Gemma 4 31B measured
τ²-Bench Telecom65.5%95.6%Qwen3.5 397B A17B
APEX-AgentsNot measured24.9%Only Qwen3.5 397B A17B measured
Neither model has a result on the other 33 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index44
Reasoning benchmarks for Gemma 4 31B and Qwen3.5 397B A17B
BenchmarkGemma 4 31BQwen3.5 397B A17BLeader
LiveBench reasoning68.3%Not measuredOnly Gemma 4 31B measured
GPQA Diamond (Artificial Analysis run)85.7%89.3%Qwen3.5 397B A17B
Humanity's Last Exam (Artificial Analysis run)23.6%29.0%Qwen3.5 397B A17B
GPQA Diamond (Epoch AI run)75.8%86.4%Qwen3.5 397B A17B
Mystery game puzzlesNot measured18.0%Only Qwen3.5 397B A17B measured
LMArena Text: hard prompts14481448Qwen3.5 397B A17B
LMArena Text: expert14641462Gemma 4 31B
LMArena Text: hard prompts in English14561450Gemma 4 31B
LMArena Vision: diagrams12831277Gemma 4 31B
LMArena Vision: homework13101292Gemma 4 31B
Neither model has a result on the other 11 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index21
Math benchmarks for Gemma 4 31B and Qwen3.5 397B A17B
BenchmarkGemma 4 31BQwen3.5 397B A17BLeader
OTIS Mock AIME73.3%88.9%Qwen3.5 397B A17B
FrontierMath, tiers 1 to 3Not measured31.2%Only Qwen3.5 397B A17B measured
LMArena Text: math14631453Gemma 4 31B
LMArena Text: mathematical work14651454Gemma 4 31B
Neither model has a result on the other 9 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index13
Knowledge benchmarks for Gemma 4 31B and Qwen3.5 397B A17B
BenchmarkGemma 4 31BQwen3.5 397B A17BLeader
CaseLaw (Vals AI)52.6%Not measuredOnly Gemma 4 31B measured
MortgageTax (Vals AI)61.4%Not measuredOnly Gemma 4 31B measured
SimpleQA Verified10.4%Not measuredOnly Gemma 4 31B measured
LMArena Text: science14571457Gemma 4 31B
LMArena Text: medicine and healthcare14511455Qwen3.5 397B A17B
LMArena Text: legal and government14381441Qwen3.5 397B A17B
LMArena Text: business and finance14301435Qwen3.5 397B A17B
Neither model has a result on the other 10 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
Web design benchmarks for Gemma 4 31B and Qwen3.5 397B A17B
BenchmarkGemma 4 31BQwen3.5 397B A17BLeader
Design Arena: websitesNot measured1191Only Qwen3.5 397B A17B measured
Design Arena: UI componentsNot measured1163Only Qwen3.5 397B A17B measured
Design Arena: game devNot measured1141Only Qwen3.5 397B A17B measured
Design Arena: data vizNot measured1175Only Qwen3.5 397B A17B measured
Design Arena: 3DNot measured1161Only Qwen3.5 397B A17B measured
Design Arena: all code categoriesNot measured1180Only Qwen3.5 397B A17B measured
Design Arena: SVGNot measured1139Only Qwen3.5 397B A17B measured
LiveBench design58.4%Not measuredOnly Gemma 4 31B measured
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index188
General benchmarks for Gemma 4 31B and Qwen3.5 397B A17B
BenchmarkGemma 4 31BQwen3.5 397B A17BLeader
LiveBench data analysis53.3%Not measuredOnly Gemma 4 31B measured
SAGE (Vals AI)55.0%Not measuredOnly Gemma 4 31B measured
IFBench75.6%78.8%Qwen3.5 397B A17B
Long context reasoning (AA-LCR)69.7%77.3%Qwen3.5 397B A17B
DTBench82.7%87.5%Qwen3.5 397B A17B
LMCA39.3%37.9%Gemma 4 31B
Surface Evolver Bench30.6%Not measuredOnly Gemma 4 31B measured
LMArena Text14431437Gemma 4 31B
LMArena Text: instruction following14331424Gemma 4 31B
LMArena Text: creative writing14171402Gemma 4 31B
LMArena Vision12761262Gemma 4 31B
LMArena Text: English14511443Gemma 4 31B
LMArena Text: non-English14301430Gemma 4 31B
LMArena Text: Chinese14751498Qwen3.5 397B A17B
LMArena Text: GermanNot measured1445Only Qwen3.5 397B A17B measured
LMArena Text: JapaneseNot measured1429Only Qwen3.5 397B A17B measured
LMArena Text: KoreanNot measured1384Only Qwen3.5 397B A17B measured
LMArena Text: French14361462Qwen3.5 397B A17B
LMArena Text: Spanish14421441Gemma 4 31B
LMArena Text: Russian14601430Gemma 4 31B
LMArena Text: Polish14331447Qwen3.5 397B A17B
LMArena Text: multi-turn14521446Gemma 4 31B
LMArena Text: longer queries14461441Gemma 4 31B
LMArena Text: writing and language14321415Gemma 4 31B
LMArena Text: entertainment and media14051399Gemma 4 31B
LMArena Vision: English12711268Gemma 4 31B
LMArena Vision: Chinese13261314Gemma 4 31B
LMArena Vision: reading text12821271Gemma 4 31B
LMArena Vision: humor12801262Gemma 4 31B
LMArena Vision: captioning12181230Qwen3.5 397B A17B
LMArena Vision: recognizing things12261265Qwen3.5 397B A17B
LMArena Vision: creative writing12851260Gemma 4 31B
LMArena Document1425Not measuredOnly Gemma 4 31B measured
Neither model has a result on the other 13 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Gemma 4 31B and Qwen3.5 397B A17B
BenchmarkGemma 4 31BQwen3.5 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

Gemma 4 31B$0.0004Fits in one request
Qwen3.5 397B A17B$0.0016Fits in one request

Gemma 4 31B costs less

Repository review

50K input + 3K output tokens

Gemma 4 31B$0.011Fits in one request
Qwen3.5 397B A17B$0.027Fits in one request

Gemma 4 31B costs less

Agent loop

200K input + 10K output tokens

Gemma 4 31B$0.044Fits in one request
Qwen3.5 397B A17B$0.101Fits in one request

Gemma 4 31B 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 Gemma 4 31B and Qwen3.5 397B A17B
SpecificationGoogle Gemma 4 31BQwen3.5 397B A17B
MakerGoogleQwen
WeightsOpen weightsOpen weights
ReleasedApril 2, 2026February 16, 2026
Context window262K262K
Input $/Mtok$0.20$0.39
Output $/Mtok$0.40$2.34
Coding index38.328.6
Independent measurements behind it21
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Gemma 4 31B better than Qwen3.5 397B A17B for coding?

Pick Gemma 4 31B for coding: it scores 38.3 on the Vibe Coding Index against 28.6 for Qwen3.5 397B A17B. Pick Qwen3.5 397B A17B when what matters is LMArena WebDev or Terminal-Bench Hard (Artificial Analysis run), where it leads.

Which is cheaper, Gemma 4 31B or Qwen3.5 397B A17B?

Gemma 4 31B costs $0.20 per million input tokens and $0.40 per million output tokens; Qwen3.5 397B A17B costs $0.39 and $2.34. Cost per solved task is known for neither of them.

Which has the bigger context window, Gemma 4 31B or Qwen3.5 397B A17B?

Neither: both take 262K tokens.

How many benchmarks is the Gemma 4 31B vs Qwen3.5 397B A17B comparison based on?

52 benchmarks have a result for both models as of October 21, 2026. Gemma 4 31B's coding index rests on 2 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