Enter opens the highlighted result. Esc closes.

Gemma 4 31B vs Kimi K2.7 Code

Kimi K2.7 Code leads Gemma 4 31B on coding, 53.4 to 39.0 on the Vibe Coding Index, and Gemma 4 31B costs 10.0x less per output token ($0.40 vs $4.00 per million). Updated October 21, 2026; 20 benchmarks have a result for both.

Replace Gemma 4 31B Replace Kimi K2.7 Code
Gemma 4 31BGoogle
Coding
39.0
Overall
42.1
Output $/M
$0.40
2
Kimi K2.7 CodeMoonshot AI
Coding
53.4
Overall
56.7
Output $/M
$4.00
18
vs boards won of 20 shared

Gemma 4 31B or Kimi K2.7 Code: which to pick

Pick Kimi K2.7 Code for coding: it scores 53.4 on the Vibe Coding Index against 39.0 for Gemma 4 31B.

Pick Gemma 4 31B for output tokens 10.0x cheaper ($0.40 vs $4.00 per million) and work measured by ALE-Bench or IFBench, where it leads.

Model AGemma 4 31BGoogle39.0/1002 independent benchmarksNo cost per solved taskModel BKimi K2.7 CodeMoonshot AI53.4/1004 independent benchmarks$0.35 per solved task
Shared results
20
Gemma 4 31B only
51
Kimi K2.7 Code only
31
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 K2.7 Code 53.4 from 4 benchmarks.
Widest category
Knowledge, 26.1 points apart on average over 1 shared benchmark.
Widest single gap
SimpleQA Verified: Gemma 4 31B 10.4% against Kimi K2.7 Code 36.5%, 26.1 points apart on a 0 to 100 scale.
Output price
Gemma 4 31B $0.40 against Kimi K2.7 Code $4.00 per million output tokens, 10.0x 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 K2.7 CodeTies
Coding17
Intelligence02
Agentic work02
Reasoning03
Math01
Knowledge01
General12
All shared boards218
CodingLike-for-likeSix of these boards feed the Vibe Coding Index17
Coding benchmarks for Gemma 4 31B and Kimi K2.7 Code
BenchmarkGemma 4 31BKimi K2.7 CodeLeader
LiveBench coding49.0%Not measuredOnly Gemma 4 31B measured
LMArena WebDev13671474Kimi K2.7 Code
AA Coding Index43.460.8Kimi K2.7 Code
LiveCodeBench (Vals AI)Not measured82.0%Only Kimi K2.7 Code measured
SWE-bench Verified (Vals AI)Not measured78.2%Only Kimi K2.7 Code measured
Vibe Code Bench (Vals AI)Not measured47.2%Only Kimi K2.7 Code measured
Code migration (Vals AI)Not measured25.4%Only Kimi K2.7 Code measured
ProgramBench (Vals AI)Not measured0.0%Only Kimi K2.7 Code measured
Terminal-Bench 2.1 (Vals AI)Not measured67.0%Only Kimi K2.7 Code measured
SkillsBench (Vals AI)Not measured50.0%Only Kimi K2.7 Code measured
SciCode45.5%47.8%Kimi K2.7 Code
Terminal-Bench Hard (Artificial Analysis run)36.4%44.7%Kimi K2.7 Code
FrontierCodeNot measured30.1%Only Kimi K2.7 Code measured
WeirdML52.3%54.1%Kimi K2.7 Code
ALE-Bench925.5886.2Gemma 4 31B
LMArena Text: coding1459Not measuredOnly Gemma 4 31B measured
LMArena Text: software and IT1463Not measuredOnly Gemma 4 31B measured
LMArena WebDev: React13591475Kimi K2.7 Code
LMArena WebDev: HTML13801451Kimi K2.7 Code
LMArena WebDev: from an imageNot measured1513Only Kimi K2.7 Code 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 K2.7 Code
BenchmarkGemma 4 31BKimi K2.7 CodeLeader
AA Intelligence Index14.725.8Kimi K2.7 Code
Epoch Capabilities Index143.0150.1Kimi K2.7 Code
Agentic workLike-for-likePublic boards, shown as published, kept out of the index02
Agentic work benchmarks for Gemma 4 31B and Kimi K2.7 Code
BenchmarkGemma 4 31BKimi K2.7 CodeLeader
AA Agentic Index4.221.0Kimi K2.7 Code
Design Arena: agents building web appsNot measured1176Only Kimi K2.7 Code measured
Design Arena: agents building full-stack appsNot measured1175Only Kimi K2.7 Code measured
Design Arena: agents building mobile appsNot measured1143Only Kimi K2.7 Code measured
Design Arena: agents building Android appsNot measured1137Only Kimi K2.7 Code measured
Design Arena: agents building Godot gamesNot measured1147Only Kimi K2.7 Code measured
Design Arena: agents building gamesNot measured1131Only Kimi K2.7 Code measured
Design Arena: HTML slidesNot measured1201Only Kimi K2.7 Code measured
Design Arena: PowerPoint slidesNot measured1154Only Kimi K2.7 Code 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%90.1%Kimi K2.7 Code
APEX-AgentsNot measured37.6%Only Kimi K2.7 Code measured
Vending-Bench 2Not measured$5,083Only Kimi K2.7 Code measured
Neither model has a result on the other 26 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index03
Reasoning benchmarks for Gemma 4 31B and Kimi K2.7 Code
BenchmarkGemma 4 31BKimi K2.7 CodeLeader
LiveBench reasoning68.3%Not measuredOnly Gemma 4 31B measured
GPQA Diamond (Artificial Analysis run)85.7%89.6%Kimi K2.7 Code
Humanity's Last Exam (Artificial Analysis run)23.6%35.0%Kimi K2.7 Code
GPQA Diamond (Epoch AI run)75.8%87.9%Kimi K2.7 Code
SimpleBenchNot measured57.9%Only Kimi K2.7 Code measured
Chess puzzlesNot measured14.0%Only Kimi K2.7 Code measured
LMArena Text: hard prompts1448Not measuredOnly Gemma 4 31B measured
LMArena Text: expert1464Not measuredOnly Gemma 4 31B measured
LMArena Text: hard prompts in English1456Not measuredOnly Gemma 4 31B measured
LMArena Vision: diagrams1283Not measuredOnly Gemma 4 31B measured
LMArena Vision: homework1310Not measuredOnly Gemma 4 31B measured
Kagi LLM Benchmark63.5%Not measuredOnly Gemma 4 31B measured
Neither model has a result on the other 10 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index01
Math benchmarks for Gemma 4 31B and Kimi K2.7 Code
BenchmarkGemma 4 31BKimi K2.7 CodeLeader
OTIS Mock AIME73.3%95.6%Kimi K2.7 Code
FrontierMath, tiers 1 to 3Not measured54.0%Only Kimi K2.7 Code measured
FrontierMath, tier 4Not measured12.2%Only Kimi K2.7 Code measured
LMArena Text: math1463Not measuredOnly Gemma 4 31B measured
LMArena Text: mathematical work1465Not measuredOnly Gemma 4 31B 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 Gemma 4 31B and Kimi K2.7 Code
BenchmarkGemma 4 31BKimi K2.7 CodeLeader
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%36.5%Kimi K2.7 Code
LMArena Text: science1457Not measuredOnly Gemma 4 31B measured
LMArena Text: medicine and healthcare1451Not measuredOnly Gemma 4 31B measured
LMArena Text: legal and government1438Not measuredOnly Gemma 4 31B measured
LMArena Text: business and finance1430Not measuredOnly Gemma 4 31B measured
FACTS Benchmark Suite45.3%Not measuredOnly Gemma 4 31B measured
Neither model has a result on the other 10 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 K2.7 Code
BenchmarkGemma 4 31BKimi K2.7 CodeLeader
Design Arena: websitesNot measured1264Only Kimi K2.7 Code measured
Design Arena: UI componentsNot measured1259Only Kimi K2.7 Code measured
Design Arena: game devNot measured1220Only Kimi K2.7 Code measured
Design Arena: data vizNot measured1227Only Kimi K2.7 Code measured
Design Arena: 3DNot measured1241Only Kimi K2.7 Code measured
Design Arena: all code categoriesNot measured1253Only Kimi K2.7 Code measured
Design Arena: SVGNot measured1176Only Kimi K2.7 Code measured
Design Arena: ASCII artNot measured1196Only Kimi K2.7 Code measured
LiveBench design58.4%Not measuredOnly Gemma 4 31B measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index12
General benchmarks for Gemma 4 31B and Kimi K2.7 Code
BenchmarkGemma 4 31BKimi K2.7 CodeLeader
LiveBench data analysis53.3%Not measuredOnly Gemma 4 31B measured
SAGE (Vals AI)55.0%Not measuredOnly Gemma 4 31B measured
IFBench75.6%63.1%Gemma 4 31B
Long context reasoning (AA-LCR)69.7%79.3%Kimi K2.7 Code
DTBench82.7%Not measuredOnly Gemma 4 31B measured
LMCA39.3%Not measuredOnly Gemma 4 31B measured
Surface Evolver Bench30.6%48.8%Kimi K2.7 Code
LMArena Text1443Not measuredOnly Gemma 4 31B measured
LMArena Text: instruction following1433Not measuredOnly Gemma 4 31B measured
LMArena Text: creative writing1417Not measuredOnly Gemma 4 31B measured
LMArena Vision1276Not measuredOnly Gemma 4 31B measured
LMArena Text: English1451Not measuredOnly Gemma 4 31B measured
LMArena Text: non-English1430Not measuredOnly Gemma 4 31B measured
LMArena Text: Chinese1475Not measuredOnly Gemma 4 31B measured
LMArena Text: French1436Not measuredOnly Gemma 4 31B measured
LMArena Text: Spanish1442Not measuredOnly Gemma 4 31B measured
LMArena Text: Russian1460Not measuredOnly Gemma 4 31B measured
LMArena Text: Polish1433Not measuredOnly Gemma 4 31B measured
LMArena Text: multi-turn1452Not measuredOnly Gemma 4 31B measured
LMArena Text: longer queries1446Not measuredOnly Gemma 4 31B measured
LMArena Text: writing and language1432Not measuredOnly Gemma 4 31B measured
LMArena Text: entertainment and media1405Not measuredOnly Gemma 4 31B measured
LMArena Vision: English1271Not measuredOnly Gemma 4 31B measured
LMArena Vision: Chinese1326Not measuredOnly Gemma 4 31B measured
LMArena Vision: reading text1282Not measuredOnly Gemma 4 31B measured
LMArena Vision: humor1280Not measuredOnly Gemma 4 31B measured
LMArena Vision: captioning1218Not measuredOnly Gemma 4 31B measured
LMArena Vision: recognizing things1226Not measuredOnly Gemma 4 31B measured
LMArena Vision: creative writing1285Not measuredOnly Gemma 4 31B measured
LMArena Document1425Not measuredOnly Gemma 4 31B 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 Gemma 4 31B and Kimi K2.7 Code
BenchmarkGemma 4 31BKimi K2.7 CodeLeader
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 K2.7 Code$0.0029Fits in one request

Gemma 4 31B costs less

Repository review

50K input + 3K output tokens

Gemma 4 31B$0.0082Fits in one request
Kimi K2.7 Code$0.059Fits in one request

Gemma 4 31B costs less

Agent loop

200K input + 10K output tokens

Gemma 4 31B$0.032Fits in one request
Kimi K2.7 Code$0.230Fits 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 K2.7 Code
SpecificationGoogle Gemma 4 31BMoonshotAI Kimi K2.7 Code
MakerGoogleMoonshot AI
WeightsOpen weightsOpen weights
ReleasedApril 2, 2026June 12, 2026
Context window262K262K
Input $/Mtok$0.14$0.95
Output $/Mtok$0.40$4.00
Coding index39.053.4
Independent measurements behind it24
Cost per solved taskNot computable$0.35
SourceOpenRouter listingOpenRouter listing

Questions

Is Gemma 4 31B better than Kimi K2.7 Code for coding?

Pick Kimi K2.7 Code for coding: it scores 53.4 on the Vibe Coding Index against 39.0 for Gemma 4 31B. Pick Gemma 4 31B for output tokens 10.0x cheaper ($0.40 vs $4.00 per million) and work measured by ALE-Bench or IFBench, where it leads.

Which is cheaper, Gemma 4 31B or Kimi K2.7 Code?

Gemma 4 31B costs $0.14 per million input tokens and $0.40 per million output tokens; Kimi K2.7 Code costs $0.95 and $4.00. Cost per solved task is known for Kimi K2.7 Code only.

Which has the bigger context window, Gemma 4 31B or Kimi K2.7 Code?

Neither: both take 262K tokens.

How many benchmarks is the Gemma 4 31B vs Kimi K2.7 Code comparison based on?

20 benchmarks have a result for both models as of October 21, 2026. Gemma 4 31B's coding index rests on 2 independent benchmarks and Kimi K2.7 Code'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/gemma-4-31b-it-vs-kimi-k2-7-code" title="Gemma 4 31B vs Kimi K2.7 Code, 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-k2-7-code">LLMBenchmark.ai, Gemma 4 31B vs Kimi K2.7 Code</a></p>