Enter opens the highlighted result. Esc closes.

Gemma 4 26B A4B vs Phi 4

Gemma 4 26B A4B leads Phi 4 on the Artificial Analysis Intelligence Index, 16.7 to 5.9, and Phi 4 costs 4.3x less per output token ($0.14 vs $0.60 per million). Updated October 21, 2026; 34 benchmarks have a result for both.

Replace Gemma 4 26B A4B Replace Phi 4
Gemma 4 26B A4BGoogle
Coding
No data
Overall
16.7
Output $/M
$0.60
34
Phi 4Microsoft
Coding
No data
Overall
5.9
Output $/M
$0.14
0
vs boards won of 34 shared

Gemma 4 26B A4B or Phi 4: which to pick

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

Pick Gemma 4 26B A4B for a larger context window (262K against 16K tokens).

Model AGemma 4 26B A4BGoogleNot benchmarked1 independent benchmarkNo cost per solved taskModel BPhi 4MicrosoftNot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
34
Gemma 4 26B A4B only
20
Phi 4 only
11
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 26B A4B not benchmarked, Phi 4 not benchmarked.
Widest category
Math, 47.6 points apart on average over 3 shared benchmarks.
Widest single gap
OTIS Mock AIME: Gemma 4 26B A4B 82.2% against Phi 4 13.8%, 68.5 points apart on a 0 to 100 scale.
Output price
Gemma 4 26B A4B $0.60 against Phi 4 $0.14 per million output tokens, 4.3x apart.
Context window
Gemma 4 26B A4B takes 262K tokens, Phi 4 16K.

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 26B A4BBarsPhi 4Ties
Coding30
Intelligence20
Agentic work10
Reasoning60
Math30
Knowledge40
General150
All shared boards340
CodingLike-for-likeSix of these boards feed the Vibe Coding Index30
Coding benchmarks for Gemma 4 26B A4B and Phi 4
BenchmarkGemma 4 26B A4BPhi 4Leader
LMArena WebDev1360Not measuredOnly Gemma 4 26B A4B measured
AA Coding Index39.3Not measuredOnly Gemma 4 26B A4B measured
LiveCodeBench (Artificial Analysis run)Not measured23.1%Only Phi 4 measured
SciCode40.0%Not measuredOnly Gemma 4 26B A4B measured
Terminal-Bench Hard (Artificial Analysis run)25.0%3.8%Gemma 4 26B A4B
WeirdML35.2%Not measuredOnly Gemma 4 26B A4B measured
ALE-Bench927.2Not measuredOnly Gemma 4 26B A4B measured
LMArena Text: coding14461231Gemma 4 26B A4B
LMArena Text: software and IT14541228Gemma 4 26B A4B
LMArena WebDev: React1351Not measuredOnly Gemma 4 26B A4B measured
LMArena WebDev: HTML1362Not measuredOnly Gemma 4 26B A4B measured
Neither model has a result on the other 27 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Gemma 4 26B A4B and Phi 4
BenchmarkGemma 4 26B A4BPhi 4Leader
AA Intelligence Index16.75.9Gemma 4 26B A4B
Epoch Capabilities Index141.9130.5Gemma 4 26B A4B
Agentic workLike-for-likePublic boards, shown as published, kept out of the index10
Agentic work benchmarks for Gemma 4 26B A4B and Phi 4
BenchmarkGemma 4 26B A4BPhi 4Leader
AA Agentic Index2.9Not measuredOnly Gemma 4 26B A4B measured
τ²-Bench Telecom43.6%0.0%Gemma 4 26B A4B
BALROGNot measured11.6%Only Phi 4 measured
Neither model has a result on the other 36 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index60
Reasoning benchmarks for Gemma 4 26B A4B and Phi 4
BenchmarkGemma 4 26B A4BPhi 4Leader
GPQA Diamond (Artificial Analysis run)79.2%57.5%Gemma 4 26B A4B
Humanity's Last Exam (Artificial Analysis run)19.3%3.8%Gemma 4 26B A4B
GPQA Diamond (Epoch AI run)73.2%56.1%Gemma 4 26B A4B
LMArena Text: hard prompts14381220Gemma 4 26B A4B
LMArena Text: expert14471203Gemma 4 26B A4B
LMArena Text: hard prompts in English14481230Gemma 4 26B A4B
LMArena Vision: diagrams1264Not measuredOnly Gemma 4 26B A4B measured
LMArena Vision: homework1293Not measuredOnly Gemma 4 26B A4B measured
Neither model has a result on the other 13 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index30
Math benchmarks for Gemma 4 26B A4B and Phi 4
BenchmarkGemma 4 26B A4BPhi 4Leader
AA Math IndexNot measured18.0Only Phi 4 measured
AIME 2025 (Artificial Analysis run)Not measured18.0%Only Phi 4 measured
OTIS Mock AIME82.2%13.8%Gemma 4 26B A4B
MATH level 5Not measured64.9%Only Phi 4 measured
LMArena Text: math14701246Gemma 4 26B A4B
LMArena Text: mathematical work14751254Gemma 4 26B A4B
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index40
Knowledge benchmarks for Gemma 4 26B A4B and Phi 4
BenchmarkGemma 4 26B A4BPhi 4Leader
MMLU Pro (Artificial Analysis run)Not measured71.4%Only Phi 4 measured
MMLUNot measured84.8%Only Phi 4 measured
LMArena Text: science14461206Gemma 4 26B A4B
LMArena Text: medicine and healthcare14281202Gemma 4 26B A4B
LMArena Text: legal and government14571241Gemma 4 26B A4B
LMArena Text: business and finance14291199Gemma 4 26B A4B
Neither model has a result on the other 11 boards here
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Web design benchmarks for Gemma 4 26B A4B and Phi 4
BenchmarkGemma 4 26B A4BPhi 4Leader
Neither model has a result on 9 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index150
General benchmarks for Gemma 4 26B A4B and Phi 4
BenchmarkGemma 4 26B A4BPhi 4Leader
IFBench72.4%23.5%Gemma 4 26B A4B
Long context reasoning (AA-LCR)65.7%0.0%Gemma 4 26B A4B
DTBench74.9%Not measuredOnly Gemma 4 26B A4B measured
LMCA29.7%Not measuredOnly Gemma 4 26B A4B measured
Creative writing (Lech Mazur)Not measured62.6%Only Phi 4 measured
LMArena Text14341217Gemma 4 26B A4B
LMArena Text: instruction following14201201Gemma 4 26B A4B
LMArena Text: creative writing14001181Gemma 4 26B A4B
LMArena Vision1260Not measuredOnly Gemma 4 26B A4B measured
LMArena Text: English14441231Gemma 4 26B A4B
LMArena Text: non-English14211197Gemma 4 26B A4B
LMArena Text: Chinese14951212Gemma 4 26B A4B
LMArena Text: GermanNot measured1222Only Phi 4 measured
LMArena Text: JapaneseNot measured1158Only Phi 4 measured
LMArena Text: KoreanNot measured1151Only Phi 4 measured
LMArena Text: French14601224Gemma 4 26B A4B
LMArena Text: Spanish14151234Gemma 4 26B A4B
LMArena Text: Russian14341209Gemma 4 26B A4B
LMArena Text: multi-turn14411206Gemma 4 26B A4B
LMArena Text: longer queries14271217Gemma 4 26B A4B
LMArena Text: writing and language14051198Gemma 4 26B A4B
LMArena Text: entertainment and media14031162Gemma 4 26B A4B
LMArena Vision: English1256Not measuredOnly Gemma 4 26B A4B measured
LMArena Vision: Chinese1312Not measuredOnly Gemma 4 26B A4B measured
LMArena Vision: reading text1268Not measuredOnly Gemma 4 26B A4B measured
LMArena Vision: humor1272Not measuredOnly Gemma 4 26B A4B measured
LMArena Vision: captioning1297Not measuredOnly Gemma 4 26B A4B measured
LMArena Vision: recognizing things1191Not measuredOnly Gemma 4 26B A4B measured
LMArena Vision: creative writing1256Not measuredOnly Gemma 4 26B A4B measured
Neither model has a result on the other 17 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Gemma 4 26B A4B and Phi 4
BenchmarkGemma 4 26B A4BPhi 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 4 26B A4B$0.0004Fits in one request
Phi 4$0.0001Fits in one request

Phi 4 costs less

Repository review

50K input + 3K output tokens

Gemma 4 26B A4B$0.0093Fits in one request
Phi 4$0.0039Exceeds the context window

Phi 4 costs less

Agent loop

200K input + 10K output tokens

Gemma 4 26B A4B$0.036Fits in one request
Phi 4$0.015Exceeds the context window

Phi 4 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 26B A4B and Phi 4
SpecificationGoogle Gemma 4 26B A4BMicrosoft Phi 4
MakerGoogleMicrosoft
WeightsOpen weightsOpen weights
ReleasedApril 3, 2026January 10, 2025
Context window262K16K
Input $/Mtok$0.15$0.07
Output $/Mtok$0.60$0.14
Coding indexNot benchmarkedNot benchmarked
Independent measurements behind it10
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Gemma 4 26B A4B better than Phi 4 for coding?

Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Gemma 4 26B A4B for a larger context window (262K against 16K tokens).

Which is cheaper, Gemma 4 26B A4B or Phi 4?

Gemma 4 26B A4B costs $0.15 per million input tokens and $0.60 per million output tokens; Phi 4 costs $0.07 and $0.14. Cost per solved task is known for neither of them.

Which has the bigger context window, Gemma 4 26B A4B or Phi 4?

Gemma 4 26B A4B, with 262K tokens against 16K.

How many benchmarks is the Gemma 4 26B A4B vs Phi 4 comparison based on?

34 benchmarks have a result for both models as of October 21, 2026. Gemma 4 26B A4B's coding index rests on 1 independent benchmark and Phi 4's on 0. Where only one model has a result, the page says so instead of naming a winner.

Compare next