Gemma 4 26B A4B
Google · 2 thinking configurations
Intelligence16.7
Input / 1M$0.10
Output / 1M$0.37
Context256K
Tariff: Google · Source: artificial-analysis · USD per million tokens
Thinking configuration
What does more thinking buy you?
Compare this model’s measured thinking configurations.
Scroll the chart ↔ · tap a point for details
Efficient frontier: best measured score at each cost2 measured configurations · 0 without both values
Selected: thinking thinking · Click a point to switch configuration
Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind. Despite 25.2B total parameters, only 3.8B activate per token during inference — delivering near-31B quality at...
Intelligence IndexArtificial Analysis composite index, not a percentage. Different versions of the index are not directly comparable.
16.7 pointsGPQA DiamondGraduate-level scientific questions. Measures specialist reasoning rather than web-search quality.
79.2 %Humanity’s Last ExamDifficult questions across many disciplines, compared under the same evaluation setup.
19.3 %SciCodeCode generation for scientific research problems.
40.0 %Terminal-Bench 2.1Agent tasks in a terminal environment. Results also depend on the evaluation harness.
39.0 %τ²-BenchTool use and interaction in assistance scenarios. Does not replace an evaluation of your integration.
43.6 %IFBenchVerifiable instruction following. A proxy for precision rather than a benchmark of creative style.
72.4 %MMMU-ProMultimodal understanding of academic problems that include images.
69.2 %Omniscience accuracyFactual accuracy in Artificial Analysis's Omniscience evaluation.
19.1 %Specifications from the catalog
Inputimage, text, video
Outputtext
Tool callingYes
Structured outputYes
SpeedNot available
First tokenNot available
Maximum output235.9K
Catalog listing date2026-04-03