Enter opens the highlighted result. Esc closes.

Gemma 3 27B vs GPT-4

Updated October 21, 2026. 36 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace Gemma 3 27B Replace GPT-4
Gemma 3 27BGoogle
Coding
4.9
Overall
4.9
Output $/M
$0.35
31
GPT-4OpenAI
Coding
22.4
Overall
6.7
Output $/M
$60.00
5
vs boards won of 36 shared

Decision reading

Pick GPT-4 for coding: it scores 22.4 on the Vibe Coding Index against 4.9 for Gemma 3 27B.

Pick Gemma 3 27B when output tokens cost 171.4x less, or when the context window is larger, 131K against 8K tokens.

Model AGemma 3 27BGoogle4.9/1001 independent benchmarkNo cost per solved taskModel BGPT-4OpenAI22.4/1001 independent benchmarkNo cost per solved task
Shared results
36
Gemma 3 27B only
26
GPT-4 only
9
Like-for-like categories
6 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 3 27B 4.9 from 1 benchmark, GPT-4 22.4 from 1 benchmark.
Widest category
Knowledge, 30.5 points apart on average over 5 shared benchmarks.
Widest single gap
LMArena Text: medicine and healthcare: Gemma 3 27B 1391 against GPT-4 1145, 41.0 points apart on a 0 to 100 scale.
Output price
Gemma 3 27B $0.35 against GPT-4 $60.00 per million output tokens, 171.4x apart.
Context window
Gemma 3 27B takes 131K tokens, GPT-4 8K.

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 3 27BBarsGPT-4Ties
Coding21
Intelligence11
Reasoning40
Math20
Knowledge50
General173
All shared boards315
CodingLike-for-likeFour of these boards feed the Vibe Coding Index21
Coding benchmarks for Gemma 3 27B and GPT-4
BenchmarkGemma 3 27BGPT-4Leader
Aider polyglot4.9%Not measuredOnly Gemma 3 27B measured
SWE-bench VerifiedNot measured22.4%Only GPT-4 measured
SWE-bench LiteNot measured39.7%Only GPT-4 measured
SWE-bench (full test set)Not measured12.5%Only GPT-4 measured
AA Coding Index10.113.1GPT-4
LiveCodeBench (Artificial Analysis run)13.7%Not measuredOnly Gemma 3 27B measured
SciCode23.3%Not measuredOnly Gemma 3 27B measured
Terminal-Bench Hard (Artificial Analysis run)3.8%Not measuredOnly Gemma 3 27B measured
LMArena Text: coding13221209Gemma 3 27B
LMArena Text: software and IT13541195Gemma 3 27B
Neither model has a result on the other 28 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index11
Intelligence benchmarks for Gemma 3 27B and GPT-4
BenchmarkGemma 3 27BGPT-4Leader
AA Intelligence Index4.96.7GPT-4
Epoch Capabilities Index130.0125.7Gemma 3 27B
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
Agentic work benchmarks for Gemma 3 27B and GPT-4
BenchmarkGemma 3 27BGPT-4Leader
AA Agentic Index0.1Not measuredOnly Gemma 3 27B measured
τ²-Bench Telecom10.5%Not measuredOnly Gemma 3 27B measured
Neither model has a result on the other 37 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index40
Reasoning benchmarks for Gemma 3 27B and GPT-4
BenchmarkGemma 3 27BGPT-4Leader
GPQA Diamond (Artificial Analysis run)42.8%34.9%Gemma 3 27B
Humanity's Last Exam (Artificial Analysis run)4.4%Not measuredOnly Gemma 3 27B measured
GPQA Diamond (Epoch AI run)47.7%Not measuredOnly Gemma 3 27B measured
BIG-Bench HardNot measured75.1%Only GPT-4 measured
LMArena Text: hard prompts13401200Gemma 3 27B
LMArena Text: expert13041149Gemma 3 27B
LMArena Text: hard prompts in English13531209Gemma 3 27B
LMArena Vision: diagrams1156Not measuredOnly Gemma 3 27B measured
LMArena Vision: homework1160Not measuredOnly Gemma 3 27B measured
Neither model has a result on the other 12 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index20
Math benchmarks for Gemma 3 27B and GPT-4
BenchmarkGemma 3 27BGPT-4Leader
AA Math Index20.7Not measuredOnly Gemma 3 27B measured
AIME 2025 (Artificial Analysis run)20.7%Not measuredOnly Gemma 3 27B measured
OTIS Mock AIME22.5%Not measuredOnly Gemma 3 27B measured
MATH level 574.0%Not measuredOnly Gemma 3 27B measured
GSM8KNot measured92.0%Only GPT-4 measured
LMArena Text: math13121230Gemma 3 27B
LMArena Text: mathematical work13271228Gemma 3 27B
Neither model has a result on the other 6 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index50
Knowledge benchmarks for Gemma 3 27B and GPT-4
BenchmarkGemma 3 27BGPT-4Leader
MMLU Pro (Artificial Analysis run)66.9%56.2%Gemma 3 27B
MMLUNot measured86.4%Only GPT-4 measured
GeoBench52.0%Not measuredOnly Gemma 3 27B measured
AI2 Reasoning ChallengeNot measured96.3%Only GPT-4 measured
LMArena Text: science13831177Gemma 3 27B
LMArena Text: medicine and healthcare13911145Gemma 3 27B
LMArena Text: legal and government13751191Gemma 3 27B
LMArena Text: business and finance13601145Gemma 3 27B
Neither model has a result on the other 9 boards here
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Web design benchmarks for Gemma 3 27B and GPT-4
BenchmarkGemma 3 27BGPT-4Leader
Neither model has a result on 9 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index173
General benchmarks for Gemma 3 27B and GPT-4
BenchmarkGemma 3 27BGPT-4Leader
IFBench31.8%33.2%GPT-4
Long context reasoning (AA-LCR)7.3%Not measuredOnly Gemma 3 27B measured
DTBench52.5%62.7%GPT-4
LMCA12.3%17.1%GPT-4
Fiction.LiveBench33.3%Not measuredOnly Gemma 3 27B measured
Creative writing (Lech Mazur)79.9%Not measuredOnly Gemma 3 27B measured
HellaSwagNot measured95.3%Only GPT-4 measured
WinoGrandeNot measured87.5%Only GPT-4 measured
LMArena Text13581206Gemma 3 27B
LMArena Text: instruction following13211201Gemma 3 27B
LMArena Text: creative writing13451191Gemma 3 27B
LMArena Vision1165Not measuredOnly Gemma 3 27B measured
LMArena Text: English13771220Gemma 3 27B
LMArena Text: non-English13341194Gemma 3 27B
LMArena Text: Chinese13461184Gemma 3 27B
LMArena Text: German13621198Gemma 3 27B
LMArena Text: Japanese12871137Gemma 3 27B
LMArena Text: Korean13081087Gemma 3 27B
LMArena Text: French13681220Gemma 3 27B
LMArena Text: Spanish13491197Gemma 3 27B
LMArena Text: Russian13491195Gemma 3 27B
LMArena Text: Polish13451265Gemma 3 27B
LMArena Text: multi-turn13451206Gemma 3 27B
LMArena Text: longer queries13331190Gemma 3 27B
LMArena Text: writing and language13471219Gemma 3 27B
LMArena Text: entertainment and media13201184Gemma 3 27B
LMArena Vision: English1188Not measuredOnly Gemma 3 27B measured
LMArena Vision: Chinese1164Not measuredOnly Gemma 3 27B measured
LMArena Vision: reading text1156Not measuredOnly Gemma 3 27B measured
LMArena Vision: humor1166Not measuredOnly Gemma 3 27B measured
LMArena Vision: captioning1151Not measuredOnly Gemma 3 27B measured
LMArena Vision: recognizing things1196Not measuredOnly Gemma 3 27B measured
LMArena Vision: creative writing1177Not measuredOnly Gemma 3 27B 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 3 27B and GPT-4
BenchmarkGemma 3 27BGPT-4Leader
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 3 27B$0.0003Fits in one request
GPT-4$0.060Fits in one request

Gemma 3 27B costs less

Repository review

50K input + 3K output tokens

Gemma 3 27B$0.0050Fits in one request
GPT-4$1.68Exceeds the context window

Gemma 3 27B costs less

Agent loop

200K input + 10K output tokens

Gemma 3 27B$0.019Exceeds the context window
GPT-4$6.60Exceeds the context window

Gemma 3 27B 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 3 27B and GPT-4
SpecificationGoogle Gemma 3 27BOpenAI GPT-4
MakerGoogleOpenAI
WeightsOpen weightsNo public weights listed
ReleasedMarch 12, 2025May 28, 2023
Context window131K8K
Input $/Mtok$0.08$30.00
Output $/Mtok$0.35$60.00
Coding index4.922.4
Independent measurements behind it11
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Gemma 3 27B better than GPT-4 for coding?

Pick GPT-4 for coding: it scores 22.4 on the Vibe Coding Index against 4.9 for Gemma 3 27B. Pick Gemma 3 27B when output tokens cost 171.4x less, or when the context window is larger, 131K against 8K tokens.

Which is cheaper, Gemma 3 27B or GPT-4?

Gemma 3 27B costs $0.08 per million input tokens and $0.35 per million output tokens; GPT-4 costs $30.00 and $60.00. Cost per solved task is known for neither of them.

Which has the bigger context window, Gemma 3 27B or GPT-4?

Gemma 3 27B, with 131K tokens against 8K.

How many benchmarks is the Gemma 3 27B vs GPT-4 comparison based on?

36 benchmarks have a result for both models as of October 21, 2026. Gemma 3 27B's coding index rests on 1 independent benchmark and GPT-4's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next