Enter opens the highlighted result. Esc closes.

Gemma 4 31B vs Kimi K2.5

Gemma 4 31B leads Kimi K2.5 on coding, 39.0 to 38.6 on the Vibe Coding Index, and Gemma 4 31B costs 7.1x less per output token ($0.40 vs $2.85 per million). Updated October 22, 2026; 54 benchmarks have a result for both.

Replace Gemma 4 31B Replace Kimi K2.5
Gemma 4 31BGoogle
Coding
39.0
Overall
42.1
Output $/M
$0.40
17
Kimi K2.5Moonshot AI
Coding
38.6
Overall
44.8
Output $/M
$2.85
37
vs boards won of 54 shared

Gemma 4 31B or Kimi K2.5: which to pick

Gemma 4 31B and Kimi K2.5 are level for coding, 39.0 against 38.6 on the Vibe Coding Index, so price and context decide.

Pick Kimi K2.5 for work measured by LMArena WebDev or AA Coding Index, where it leads. Pick Gemma 4 31B for output tokens 7.1x cheaper ($0.40 vs $2.85 per million) and work measured by Terminal-Bench Hard (Artificial Analysis run) or WeirdML, where it leads.

Model AGemma 4 31BGoogle39.0/1002 independent benchmarksNo cost per solved taskModel BKimi K2.5Moonshot AI38.6/1003 independent benchmarks$0.24 per solved task
Shared results
54
Gemma 4 31B only
17
Kimi K2.5 only
45
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
Gemma 4 31B 39.0 from 2 benchmarks, Kimi K2.5 38.6 from 3 benchmarks.
Widest category
Agentic work, 30.4 points apart on average over 1 shared benchmark.
Widest single gap
τ²-Bench Telecom: Gemma 4 31B 65.5% against Kimi K2.5 95.9%, 30.4 points apart on a 0 to 100 scale.
Output price
Gemma 4 31B $0.40 against Kimi K2.5 $2.85 per million output tokens, 7.1x 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 31BBarsKimi K2.5Ties
Coding36
Intelligence02
Agentic work01
Reasoning26
Math02
Knowledge25
General1015
All shared boards1737
CodingLike-for-likeSix of these boards feed the Vibe Coding Index36
Coding benchmarks for Gemma 4 31B and Kimi K2.5
BenchmarkGemma 4 31BKimi K2.5Leader
SWE-bench VerifiedNot measured70.8%Only Kimi K2.5 measured
SWE-bench MultilingualNot measured67.3%Only Kimi K2.5 measured
LiveBench coding49.0%Not measuredOnly Gemma 4 31B measured
LMArena WebDev13671437Kimi K2.5
AA Coding Index43.446.8Kimi K2.5
LiveCodeBench (Vals AI)Not measured83.9%Only Kimi K2.5 measured
SWE-bench Verified (Vals AI)Not measured70.0%Only Kimi K2.5 measured
Vibe Code Bench (Vals AI)Not measured17.5%Only Kimi K2.5 measured
Code migration (Vals AI)Not measured7.0%Only Kimi K2.5 measured
Terminal-Bench 2.1 (Vals AI)Not measured42.0%Only Kimi K2.5 measured
SciCode45.5%Not measuredOnly Gemma 4 31B measured
Terminal-Bench Hard (Artificial Analysis run)36.4%34.8%Gemma 4 31B
Terminal-BenchNot measured43.2%Only Kimi K2.5 measured
WeirdML52.3%45.6%Gemma 4 31B
ALE-Bench925.5821.6Gemma 4 31B
LMArena Text: coding14591483Kimi K2.5
LMArena Text: software and IT14631471Kimi K2.5
LMArena WebDev: React13591429Kimi K2.5
LMArena WebDev: HTML13801426Kimi K2.5
LMArena WebDev: from an imageNot measured1439Only Kimi K2.5 measured
SciCode8.9%Not measuredOnly Gemma 4 31B measured
Neither model has a result on the other 18 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for Gemma 4 31B and Kimi K2.5
BenchmarkGemma 4 31BKimi K2.5Leader
AA Intelligence Index14.723.5Kimi K2.5
Epoch Capabilities Index143.0148.1Kimi K2.5
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
Agentic work benchmarks for Gemma 4 31B and Kimi K2.5
BenchmarkGemma 4 31BKimi K2.5Leader
AA Agentic Index4.2Not measuredOnly Gemma 4 31B measured
Design Arena: agents building web appsNot measured1136Only Kimi K2.5 measured
Design Arena: agents building full-stack appsNot measured1114Only Kimi K2.5 measured
Design Arena: agents building mobile appsNot measured1147Only Kimi K2.5 measured
Design Arena: agents building Android appsNot measured1118Only Kimi K2.5 measured
Design Arena: agents building Godot gamesNot measured1219Only Kimi K2.5 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
Finance Agent (Vals AI)Not measured35.8%Only Kimi K2.5 measured
Legal Research Bench (Vals AI)Not measured15.9%Only Kimi K2.5 measured
Excel Modeling Benchmark (Vals AI)Not measured28.5%Only Kimi K2.5 measured
τ²-Bench Telecom65.5%95.9%Kimi K2.5
OSWorldNot measured63.3%Only Kimi K2.5 measured
Vending-Bench 2Not measured$1,198Only Kimi K2.5 measured
Neither model has a result on the other 26 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index26
Reasoning benchmarks for Gemma 4 31B and Kimi K2.5
BenchmarkGemma 4 31BKimi K2.5Leader
GPQA DiamondNot measured84.1%Only Kimi K2.5 measured
LiveBench reasoning68.3%Not measuredOnly Gemma 4 31B measured
MMMU Pro (Vals AI)Not measured84.3%Only Kimi K2.5 measured
GPQA Diamond (Artificial Analysis run)85.7%87.9%Kimi K2.5
Humanity's Last Exam (Artificial Analysis run)23.6%30.7%Kimi K2.5
GPQA Diamond (Epoch AI run)75.8%Not measuredOnly Gemma 4 31B measured
ARC-AGINot measured65.3%Only Kimi K2.5 measured
ARC-AGI-2Not measured11.8%Only Kimi K2.5 measured
SimpleBenchNot measured46.8%Only Kimi K2.5 measured
LMArena Text: hard prompts14481453Kimi K2.5
LMArena Text: expert14641466Kimi K2.5
LMArena Text: hard prompts in English14561460Kimi K2.5
LMArena Vision: diagrams12831282Gemma 4 31B
LMArena Vision: homework13101297Gemma 4 31B
Kagi LLM Benchmark63.5%78.5%Kimi K2.5
Neither model has a result on the other 7 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index02
Math benchmarks for Gemma 4 31B and Kimi K2.5
BenchmarkGemma 4 31BKimi K2.5Leader
AIMENot measured95.6%Only Kimi K2.5 measured
OTIS Mock AIME73.3%Not measuredOnly Gemma 4 31B measured
LMArena Text: math14631470Kimi K2.5
LMArena Text: mathematical work14651473Kimi K2.5
Neither model has a result on the other 9 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index25
Knowledge benchmarks for Gemma 4 31B and Kimi K2.5
BenchmarkGemma 4 31BKimi K2.5Leader
MMLU ProNot measured85.9%Only Kimi K2.5 measured
CaseLaw (Vals AI)52.6%58.7%Kimi K2.5
CorpFin (Vals AI)Not measured68.3%Only Kimi K2.5 measured
TaxEval (Vals AI)Not measured74.2%Only Kimi K2.5 measured
MortgageTax (Vals AI)61.4%66.5%Kimi K2.5
MedQA (Vals AI)Not measured94.4%Only Kimi K2.5 measured
MedCode (Vals AI)Not measured39.3%Only Kimi K2.5 measured
SimpleQA Verified10.4%34.3%Kimi K2.5
LMArena Text: science14571459Kimi K2.5
LMArena Text: medicine and healthcare14511448Gemma 4 31B
LMArena Text: legal and government14381441Kimi K2.5
LMArena Text: business and finance14301427Gemma 4 31B
FACTS Benchmark Suite45.3%Not measuredOnly Gemma 4 31B measured
Neither model has a result on the other 5 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
Web design benchmarks for Gemma 4 31B and Kimi K2.5
BenchmarkGemma 4 31BKimi K2.5Leader
Design Arena: websitesNot measured1248Only Kimi K2.5 measured
Design Arena: UI componentsNot measured1234Only Kimi K2.5 measured
Design Arena: game devNot measured1206Only Kimi K2.5 measured
Design Arena: data vizNot measured1219Only Kimi K2.5 measured
Design Arena: 3DNot measured1204Only Kimi K2.5 measured
Design Arena: all code categoriesNot measured1236Only Kimi K2.5 measured
Design Arena: SVGNot measured1149Only Kimi K2.5 measured
Design Arena: ASCII artNot measured1166Only Kimi K2.5 measured
LiveBench design58.4%Not measuredOnly Gemma 4 31B measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index1015
General benchmarks for Gemma 4 31B and Kimi K2.5
BenchmarkGemma 4 31BKimi K2.5Leader
LiveBench data analysis53.3%Not measuredOnly Gemma 4 31B measured
SAGE (Vals AI)55.0%49.9%Gemma 4 31B
MedScribe (Vals AI)Not measured76.4%Only Kimi K2.5 measured
IFBench75.6%70.2%Gemma 4 31B
Long context reasoning (AA-LCR)69.7%78.0%Kimi K2.5
DTBench82.7%Not measuredOnly Gemma 4 31B measured
LMCA39.3%Not measuredOnly Gemma 4 31B measured
Fiction.LiveBenchNot measured86.1%Only Kimi K2.5 measured
CL-benchNot measured19.3%Only Kimi K2.5 measured
Surface Evolver Bench30.6%Not measuredOnly Gemma 4 31B measured
CL-bench LifeNot measured13.2%Only Kimi K2.5 measured
LMArena Text14431445Kimi K2.5
LMArena Text: instruction following14331431Gemma 4 31B
LMArena Text: creative writing14171422Kimi K2.5
LMArena Vision12761269Gemma 4 31B
LMArena Text: English14511453Kimi K2.5
LMArena Text: non-English14301433Kimi K2.5
LMArena Text: Chinese14751494Kimi K2.5
LMArena Text: GermanNot measured1441Only Kimi K2.5 measured
LMArena Text: JapaneseNot measured1423Only Kimi K2.5 measured
LMArena Text: KoreanNot measured1412Only Kimi K2.5 measured
LMArena Text: French14361454Kimi K2.5
LMArena Text: Spanish14421450Kimi K2.5
LMArena Text: Russian14601435Gemma 4 31B
LMArena Text: Polish14331445Kimi K2.5
LMArena Text: multi-turn14521444Gemma 4 31B
LMArena Text: longer queries14461445Gemma 4 31B
LMArena Text: writing and language14321425Gemma 4 31B
LMArena Text: entertainment and media14051417Kimi K2.5
LMArena Vision: English12711274Kimi K2.5
LMArena Vision: Chinese13261328Kimi K2.5
LMArena Vision: reading text12821276Gemma 4 31B
LMArena Vision: humor12801284Kimi K2.5
LMArena Vision: captioning12181257Kimi K2.5
LMArena Vision: recognizing things12261285Kimi K2.5
LMArena Vision: creative writing12851269Gemma 4 31B
LMArena Document1425Not measuredOnly Gemma 4 31B measured
Neither model has a result on the other 9 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 Kimi K2.5
BenchmarkGemma 4 31BKimi K2.5Leader
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$0.0003Fits in one request
Kimi K2.5$0.0020Fits in one request

Gemma 4 31B costs less

Repository review

50K input + 3K output tokens

Gemma 4 31B$0.0082Fits in one request
Kimi K2.5$0.035Fits in one request

Gemma 4 31B costs less

Agent loop

200K input + 10K output tokens

Gemma 4 31B$0.032Fits in one request
Kimi K2.5$0.135Fits 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 Kimi K2.5
SpecificationGoogle Gemma 4 31BMoonshotAI Kimi K2.5
MakerGoogleMoonshot AI
WeightsOpen weightsOpen weights
ReleasedApril 2, 2026January 27, 2026
Context window262K262K
Input $/Mtok$0.14$0.532
Output $/Mtok$0.40$2.85
Coding index39.038.6
Independent measurements behind it23
Cost per solved taskNot computable$0.24
SourceOpenRouter listingOpenRouter listing

Questions

Is Gemma 4 31B better than Kimi K2.5 for coding?

Gemma 4 31B and Kimi K2.5 are level for coding, 39.0 against 38.6 on the Vibe Coding Index, so price and context decide. Pick Kimi K2.5 for work measured by LMArena WebDev or AA Coding Index, where it leads. Pick Gemma 4 31B for output tokens 7.1x cheaper ($0.40 vs $2.85 per million) and work measured by Terminal-Bench Hard (Artificial Analysis run) or WeirdML, where it leads.

Which is cheaper, Gemma 4 31B or Kimi K2.5?

Gemma 4 31B costs $0.14 per million input tokens and $0.40 per million output tokens; Kimi K2.5 costs $0.532 and $2.85. Cost per solved task is known for Kimi K2.5 only.

Which has the bigger context window, Gemma 4 31B or Kimi K2.5?

Neither: both take 262K tokens.

How many benchmarks is the Gemma 4 31B vs Kimi K2.5 comparison based on?

54 benchmarks have a result for both models as of October 22, 2026. Gemma 4 31B's coding index rests on 2 independent benchmarks and Kimi K2.5's on 3. 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/gemma-4-31b-it-vs-kimi-k2-5" title="Gemma 4 31B vs Kimi K2.5, 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/gemma-4-31b-it-vs-kimi-k2-5">LLMBenchmark.ai, Gemma 4 31B vs Kimi K2.5</a></p>