Enter opens the highlighted result. Esc closes.

Gemma 3 27B vs Llama 3.1 8B Instruct

Llama 3.1 8B Instruct leads Gemma 3 27B on the Artificial Analysis Intelligence Index, 6.9 to 4.9, and Llama 3.1 8B Instruct costs 3.8x less per output token ($0.08 vs $0.30 per million). Updated October 22, 2026; 46 benchmarks have a result for both.

Replace Gemma 3 27B Replace Llama 3.1 8B Instruct
Gemma 3 27BGoogle
Coding
No data
Overall
31.4
Output $/M
$0.30
42
Llama 3.1 8B InstructMeta Llama
Coding
No data
Overall
17.9
Output $/M
$0.08
4
vs boards won of 46 shared

Gemma 3 27B or Llama 3.1 8B Instruct: which to pick

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

Pick Gemma 3 27B for work measured by AA Intelligence Index or τ²-Bench Telecom, where it leads.

Model AGemma 3 27BGoogleNot benchmarkedNo coding benchmarkNo cost per solved taskModel BLlama 3.1 8B InstructMeta LlamaNot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
46
Gemma 3 27B only
20
Llama 3.1 8B Instruct only
5
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 3 27B not benchmarked, Llama 3.1 8B Instruct not benchmarked.
Widest category
Knowledge, 29.2 points apart on average over 5 shared benchmarks.
Widest single gap
MATH level 5: Gemma 3 27B 74.0% against Llama 3.1 8B Instruct 22.9%, 51.2 points apart on a 0 to 100 scale.
Output price
Gemma 3 27B $0.30 against Llama 3.1 8B Instruct $0.08 per million output tokens, 3.8x apart.
Context window
Both take 131K 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 3 27BBarsLlama 3.1 8B InstructTies
Coding50
Intelligence11
Agentic work11
Reasoning51
Math60
Knowledge50
General191
All shared boards424
CodingLike-for-likeSix of these boards feed the Vibe Coding Index50
Coding benchmarks for Gemma 3 27B and Llama 3.1 8B Instruct
BenchmarkGemma 3 27BLlama 3.1 8B InstructLeader
Aider polyglot4.9%Not measuredOnly Gemma 3 27B measured
AA Coding Index10.15.4Gemma 3 27B
LiveCodeBench (Artificial Analysis run)13.7%11.6%Gemma 3 27B
SciCode23.3%Not measuredOnly Gemma 3 27B measured
Terminal-Bench Hard (Artificial Analysis run)3.8%0.8%Gemma 3 27B
WeirdMLNot measured1.7%Only Llama 3.1 8B Instruct measured
LMArena Text: coding13221195Gemma 3 27B
LMArena Text: software and IT13541195Gemma 3 27B
SciCode0.0%Not measuredOnly Gemma 3 27B measured
Neither model has a result on the other 30 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index11
Intelligence benchmarks for Gemma 3 27B and Llama 3.1 8B Instruct
BenchmarkGemma 3 27BLlama 3.1 8B InstructLeader
AA Intelligence Index4.96.9Llama 3.1 8B Instruct
Epoch Capabilities Index130.0116.5Gemma 3 27B
Agentic workLike-for-likePublic boards, shown as published, kept out of the index11
Agentic work benchmarks for Gemma 3 27B and Llama 3.1 8B Instruct
BenchmarkGemma 3 27BLlama 3.1 8B InstructLeader
AA Agentic Index0.1Not measuredOnly Gemma 3 27B measured
τ²-Bench Telecom10.5%16.4%Llama 3.1 8B Instruct
BALROGNot measured15.1%Only Llama 3.1 8B Instruct measured
Berkeley Function Calling Leaderboard (BFCL)29.5%25.8%Gemma 3 27B
Neither model has a result on the other 37 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index51
Reasoning benchmarks for Gemma 3 27B and Llama 3.1 8B Instruct
BenchmarkGemma 3 27BLlama 3.1 8B InstructLeader
GPQA Diamond (Artificial Analysis run)42.8%25.9%Gemma 3 27B
Humanity's Last Exam (Artificial Analysis run)4.4%5.3%Llama 3.1 8B Instruct
GPQA Diamond (Epoch AI run)47.7%27.0%Gemma 3 27B
LMArena Text: hard prompts13401175Gemma 3 27B
LMArena Text: expert13041144Gemma 3 27B
LMArena Text: hard prompts in English13531196Gemma 3 27B
LMArena Vision: diagrams1156Not measuredOnly Gemma 3 27B measured
LMArena Vision: homework1160Not measuredOnly Gemma 3 27B measured
Kagi LLM Benchmark35.1%Not measuredOnly Gemma 3 27B measured
Neither model has a result on the other 13 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index60
Math benchmarks for Gemma 3 27B and Llama 3.1 8B Instruct
BenchmarkGemma 3 27BLlama 3.1 8B InstructLeader
AA Math Index20.74.3Gemma 3 27B
AIME 2025 (Artificial Analysis run)20.7%4.3%Gemma 3 27B
OTIS Mock AIME22.5%1.7%Gemma 3 27B
MATH level 574.0%22.9%Gemma 3 27B
GSM8KNot measured82.4%Only Llama 3.1 8B Instruct measured
LMArena Text: math13121179Gemma 3 27B
LMArena Text: mathematical work13271180Gemma 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 Llama 3.1 8B Instruct
BenchmarkGemma 3 27BLlama 3.1 8B InstructLeader
MMLU Pro (Artificial Analysis run)66.9%47.6%Gemma 3 27B
MMLUNot measured56.1%Only Llama 3.1 8B Instruct measured
GeoBench52.0%Not measuredOnly Gemma 3 27B measured
LMArena Text: science13831193Gemma 3 27B
LMArena Text: medicine and healthcare13911182Gemma 3 27B
LMArena Text: legal and government13751206Gemma 3 27B
LMArena Text: business and finance13601166Gemma 3 27B
FACTS Benchmark Suite29.1%Not measuredOnly Gemma 3 27B measured
Neither model has a result on the other 10 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 Llama 3.1 8B Instruct
BenchmarkGemma 3 27BLlama 3.1 8B InstructLeader
Neither model has a result on 9 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index191
General benchmarks for Gemma 3 27B and Llama 3.1 8B Instruct
BenchmarkGemma 3 27BLlama 3.1 8B InstructLeader
IFBench31.8%28.6%Gemma 3 27B
Long context reasoning (AA-LCR)7.3%18.0%Llama 3.1 8B Instruct
DTBench52.5%50.9%Gemma 3 27B
LMCA12.3%5.4%Gemma 3 27B
Fiction.LiveBench33.3%Not measuredOnly Gemma 3 27B measured
Creative writing (Lech Mazur)79.9%Not measuredOnly Gemma 3 27B measured
PIQANot measured81.2%Only Llama 3.1 8B Instruct measured
LMArena Text13581187Gemma 3 27B
LMArena Text: instruction following13211159Gemma 3 27B
LMArena Text: creative writing13451154Gemma 3 27B
LMArena Vision1165Not measuredOnly Gemma 3 27B measured
LMArena Text: English13771218Gemma 3 27B
LMArena Text: non-English13341148Gemma 3 27B
LMArena Text: Chinese13461151Gemma 3 27B
LMArena Text: German13621144Gemma 3 27B
LMArena Text: Japanese12871061Gemma 3 27B
LMArena Text: Korean13081053Gemma 3 27B
LMArena Text: French13681178Gemma 3 27B
LMArena Text: Spanish13491169Gemma 3 27B
LMArena Text: Russian13491158Gemma 3 27B
LMArena Text: Polish1345Not measuredOnly Gemma 3 27B measured
LMArena Text: multi-turn13451172Gemma 3 27B
LMArena Text: longer queries13331181Gemma 3 27B
LMArena Text: writing and language13471162Gemma 3 27B
LMArena Text: entertainment and media13201145Gemma 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 14 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 Llama 3.1 8B Instruct
BenchmarkGemma 3 27BLlama 3.1 8B InstructLeader
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$0.0002Fits in one request
Llama 3.1 8B Instruct$0.0001Fits in one request

Llama 3.1 8B Instruct costs less

Repository review

50K input + 3K output tokens

Gemma 3 27B$0.0049Fits in one request
Llama 3.1 8B Instruct$0.0027Fits in one request

Llama 3.1 8B Instruct costs less

Agent loop

200K input + 10K output tokens

Gemma 3 27B$0.019Exceeds the context window
Llama 3.1 8B Instruct$0.011Exceeds the context window

Llama 3.1 8B Instruct 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 Llama 3.1 8B Instruct
SpecificationGoogle Gemma 3 27BMeta Llama 3.1 8B Instruct
MakerGoogleMeta Llama
WeightsOpen weightsOpen weights
ReleasedMarch 12, 2025July 23, 2024
Context window131K131K
Input $/Mtok$0.08$0.05
Output $/Mtok$0.30$0.08
Coding indexNot benchmarkedNot benchmarked
Independent measurements behind it00
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Gemma 3 27B better than Llama 3.1 8B Instruct for coding?

Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Gemma 3 27B for work measured by AA Intelligence Index or τ²-Bench Telecom, where it leads.

Which is cheaper, Gemma 3 27B or Llama 3.1 8B Instruct?

Gemma 3 27B costs $0.08 per million input tokens and $0.30 per million output tokens; Llama 3.1 8B Instruct costs $0.05 and $0.08. Cost per solved task is known for neither of them.

Which has the bigger context window, Gemma 3 27B or Llama 3.1 8B Instruct?

Neither: both take 131K tokens.

How many benchmarks is the Gemma 3 27B vs Llama 3.1 8B Instruct comparison based on?

46 benchmarks have a result for both models as of October 22, 2026. Gemma 3 27B's coding index rests on 0 independent benchmarks and Llama 3.1 8B Instruct's on 0. Where only one model has a result, the page says so instead of naming a winner.

Compare next