Skip to main content

Model profile

Gemma 4 26B A4B

GoogleCurrentReleased Apr 2, 2026
Data verified
Overall Score
57.96Public #67 of 200Verified #44 of 99
Arena Elo
1438
Eligible category ranks
2of 8
Price (1M tokens)
$0 in / $0 out
API pricing
Speed
Not listed
Context
256K

Evidence coverage

20 of 323 tracked benchmarks are published. 4 are verified and 16 provisional. 6 of 8 categories are measured.

Updated July 23, 2026Methodology
Published / tracked
20 / 323
Verified
4
Provisional
16
Categories with evidence
6 / 8

Evidence by category

  • Agentic4 benchmarks
    Reported
  • Coding2 benchmarks
    Reported
  • Reasoning2 benchmarks
    Reported
  • Knowledge9 benchmarks
    Mixed evidence
  • Math0 benchmarks
    Not measured
  • Multilingual0 benchmarks
    Not measured
  • Multimodal2 benchmarks
    Mixed evidence
  • Inst. Following1 benchmark
    Reported
Open WeightSelf-hostReasoning
Confidence:
Low
26b-a4b

Gemma 4 26B A4B ranks #67 out of 200 models on the public leaderboard with an overall score of 57.96/100. It also ranks #44 out of 99 on the verified leaderboard. While not a frontier model, it offers specific advantages depending on the use case.

Gemma 4 26B A4B is a open weight model with a 256K token context window. It uses explicit chain-of-thought reasoning, which typically improves performance on math and complex reasoning tasks at the cost of higher latency and token usage.

Gemma 4 26B A4B sits inside the Gemma 4 family alongside Gemma 4 31B, Gemma 4 12B, Gemma 4 E2B, Gemma 4 E4B. This profile currently has 20 of 323 tracked benchmarks. BenchLM only exposes non-generated benchmark rows publicly, so missing categories stay blank until a sourced evaluation is available.

Its strongest category is Agentic (#50), while its weakest is Coding (#84). This performance profile makes it particularly useful for coding agents, browser research, and computer-use workflows.

Peer position

Exact provisional scores and ranks for the closest listed peers. A score can appear before a model clears the evidence threshold for a rank, so equal scores can have different rank states.

Range 57.4458.61

  1. GPT-5 (high)
    OpenAI
    #6358.61
    GPT-5 (high) is #63 with a score of 58.61.
    Compare
  2. GPT-5.2
    OpenAI
    #6458.43
    GPT-5.2 is #64 with a score of 58.43.
    Compare
  3. DeepSeek V3.2 (Thinking)
    DeepSeek
    #6558.15
    DeepSeek V3.2 (Thinking) is #65 with a score of 58.15.
    Compare
  4. Qwen3 235B 2507 (Reasoning)
    Alibaba
    #6658.01
    Qwen3 235B 2507 (Reasoning) is #66 with a score of 58.01.
    Compare
  5. Gemma 4 26B A4BCurrent model
    Google
    #6757.96
    Gemma 4 26B A4B is #67 with a score of 57.96.
  6. GLM-4.5
    Z.AI
    #6857.56
    GLM-4.5 is #68 with a score of 57.56.
    Compare
  7. Claude Opus 4.5 Thinking
    Anthropic
    #6957.44
    Claude Opus 4.5 Thinking is #69 with a score of 57.44.
    Compare

Category percentile

More

Relative position among models eligible for each sourced category. A higher percentile means a stronger position within that category's ranked cohort; 100 is highest.

  1. Agentic58%
    Eligible cohort rank #50 of 119Category score 49.5
  2. Coding31%
    Eligible cohort rank #84 of 122Category score 46.8

Category evidence

Scores and ranks appear only where this model has published benchmark evidence. Categories without displayable source records remain not measured.

Category scores, ranks, weighting, benchmark coverage, and evidence status
CategoryScore
AgenticRank #50 of 119Percentile 58thWeight 22%4 benchmarksReported49.5
CodingRank #84 of 122Percentile 31stWeight 20%2 benchmarksReported46.8
ReasoningRank Not rankedWeight 17%2 benchmarksReported44.1
KnowledgeRank Not rankedWeight 12%9 benchmarksMixed sources38.4
MathWeight 5%0 benchmarksNot measuredNot measured
MultilingualWeight 7%0 benchmarksNot measuredNot measured
MultimodalRank Not rankedWeight 12%2 benchmarksMixed sources46.8
Inst. FollowingWeight 5%1 benchmarkReportedScore pending

Chatbot Arena performance

Scroll horizontally to inspect confidence intervals and vote counts.

Chatbot Arena Elo, confidence interval, and vote count by evaluation view
ViewEloConfidence intervalVotes
Text Overall1438±7.65,798
Coding1480±15.31,369
Math1466±28.4371
Instruction Following1439±14.11,601
Creative Writing1402±19.0946
Multi-turn1447±17.51,089
Hard Prompts1461±10.13,270
Hard Prompts (English)1468±14.81,485
Longer Query1448±14.51,557

Benchmark Details

Rows below have a displayable published verification record. Each source link and provenance note remains in the page HTML while its category is closed. Source-unverified manual rows and generated rows stay hidden.

Agentic4 benchmarks
AA Agentic IndexReported

Artificial Analysis Agentic Index

11.0%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
τ²-bench resultsReported

τ²-Bench Tool-Agent-User Evaluation

43.6%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
GDPval-AAReported

GDPval-AA normalized

13.1%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
GDPval-AAReported
761Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Coding2 benchmarks
AA Coding IndexReported

Artificial Analysis Coding Index

39.3%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-SciCodeReported

Artificial Analysis SciCode

40.0%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Reasoning2 benchmarks
AA-LCRReported

Artificial Analysis Long Context Reasoning

55.7%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
CritPtReported

Critical Physics Tasks

0.0%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Knowledge9 benchmarks
HLEProvider exact

Humanity's Last Exam

17.2%Weighted 45%
Source: Gemma 4 26B A4B model cardProvenance: Provider exact
MMLU-ProProvider exact

Massive Multitask Language Understanding Professional

82.6%Weighted 30%
Source: Gemma 4 26B A4B model cardProvenance: Provider exact
HLE w/o toolsProvider exact

Humanity's Last Exam without tools

8.7%Display only
Source: Gemma 4 26B A4B model cardProvenance: Provider exact
Artificial Analysis Intelligence IndexReported
25.7%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-GPQA DiamondReported

Artificial Analysis GPQA Diamond

79.2%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-HLEReported

Artificial Analysis Humanity's Last Exam

18.3%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience IndexReported

Artificial Analysis Omniscience Index

-48.1%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience AccuracyReported

Artificial Analysis Omniscience Accuracy

18.2%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience Hallucination RateReported

Artificial Analysis Omniscience Hallucination Rate

80.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Multimodal2 benchmarks
MMMU-ProProvider exact

Massive Multi-discipline Multimodal Understanding Pro

73.8%Weighted 45%
Source: Gemma 4 26B A4B model cardProvenance: Provider exact
AA-MMMU-ProReported

Artificial Analysis MMMU-Pro

69.2%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Inst. Following1 benchmark
AA-IFBenchReported

Artificial Analysis IFBench

72.4%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.

Frequently Asked Questions

How does Gemma 4 26B A4B perform overall in AI benchmarks?

Gemma 4 26B A4B has 20 published benchmark scores on BenchLM, but it does not yet have enough non-generated coverage to receive a global overall rank.

Is Gemma 4 26B A4B good for knowledge and understanding?

Gemma 4 26B A4B has visible benchmark coverage in knowledge and understanding, but BenchLM does not currently assign it a global category rank there.

Is Gemma 4 26B A4B good for coding and programming?

Gemma 4 26B A4B ranks #84 out of 122 models in coding and programming benchmarks with an average score of 46.8. There are stronger options in this category.

Is Gemma 4 26B A4B good for reasoning and logic?

Gemma 4 26B A4B has visible benchmark coverage in reasoning and logic, but BenchLM does not currently assign it a global category rank there.

Is Gemma 4 26B A4B good for agentic tool use and computer tasks?

Gemma 4 26B A4B ranks #50 out of 119 models in agentic tool use and computer tasks benchmarks with an average score of 49.5. There are stronger options in this category.

Is Gemma 4 26B A4B good for multimodal and grounded tasks?

Gemma 4 26B A4B has visible benchmark coverage in multimodal and grounded tasks, but BenchLM does not currently assign it a global category rank there.

Is Gemma 4 26B A4B good for instruction following?

Gemma 4 26B A4B has visible benchmark coverage in instruction following, but BenchLM does not currently assign it a global category rank there.

Is Gemma 4 26B A4B open source?

Yes, Gemma 4 26B A4B is an open weight model created by Google, meaning it can be downloaded and run locally or fine-tuned for specific use cases.

Which sibling models are related to Gemma 4 26B A4B?

Gemma 4 26B A4B belongs to the Gemma 4 family. Related variants on BenchLM include Gemma 4 31B, Gemma 4 12B, Gemma 4 E2B, Gemma 4 E4B.

Does Gemma 4 26B A4B have full benchmark coverage on BenchLM?

Not yet. Gemma 4 26B A4B currently has 20 published benchmark scores out of the 323 benchmarks BenchLM tracks. BenchLM only exposes non-generated public benchmark rows, so missing categories stay blank until a sourced evaluation is available.

What is the context window size of Gemma 4 26B A4B?

Gemma 4 26B A4B has a published context window of 256K, which determines how much text it can process in a single interaction.

Last updated: July 23, 2026 · Runtime metrics stay blank until BenchLM has a sourced snapshot.

Choose with this week’s evidence

Join 2,000+ readers for ranking moves, new releases, pricing changes, and the evidence behind them.

Free. One email per week.