What is the IQ of gemma-4-31b?

gemma-4-31b has an estimated IQ of 101, which ranks #120 out of 176 public models on the AI IQ leaderboard. The score combines seven equally weighted reasoning dimensions using measured benchmark results and conservative estimates for missing coverage.

What is gemma-4-31b best at?

Among the dimensions with published benchmark results, gemma-4-31b ranks highest in Academic Reasoning (IQ 117, #116 of 203) and Knowledge Work (IQ 103, #106 of 176). Its lowest-ranked measured dimension is Programmatic Reasoning (IQ 97, #150 of 177).

What is the EQ of gemma-4-31b?

gemma-4-31b has an estimated EQ of 105. EQ measures emotional reasoning — how well a model reads and responds to emotional context — and is scored separately from IQ.

Which benchmarks has gemma-4-31b been tested on?

gemma-4-31b has published results on 17 of the 39 benchmarks behind AI IQ's Composite IQ, including FACTS Grounding, IFBench, GPQA Diamond, SciCode and Humanity's Last Exam.

How much does gemma-4-31b cost?

gemma-4-31b costs $0.12 per million input tokens and $0.35 per million output tokens. Adjusted for how many tokens it actually consumes on real tasks, its effective cost works out to $0.1371 per million input-plus-output tokens.

Is gemma-4-31b open source?

Yes — gemma-4-31b is an open-weights model: its trained weights are published, so it can be downloaded and self-hosted.

Who developed gemma-4-31b?

gemma-4-31b was developed by Google, based in the United States.

When was gemma-4-31b released?

gemma-4-31b was publicly released on April 2, 2026.

How large is the context window of gemma-4-31b?

gemma-4-31b supports a context window of 256,000 tokens — the amount of text it can consider in a single request.

How fast is gemma-4-31b?

In independent measurements, gemma-4-31b generates about 35 tokens per second, with a typical response time of about 64 seconds.

GoogleReleased Apr 2, 2026

Open weights

gemma-4-31b

Its best results are in Academic Reasoning (#116 of 203) and Knowledge Work (#106 of 176); weakest in Programmatic Reasoning (#150 of 177). Three of seven dimensions are estimated, because no results are published yet.

Composite IQ ?

101

#120 of 176 ranked models · ahead of 32%

Google's #12 model, behind Gemini 4 Argon (140)

Effective cost

$0.137/1M

$0.12 in · $0.35 out

Speed

35tok/s

64s typical response

Context

256Ktokens

256,000 tokens in

Evidence

4of 7 dims

17 of 39 benchmarks

Compare with gpt-6.1-solAPI pricing

IQ profile

Seven dimensions, averaged equally into Composite IQ

Compare shape with

MeasuredEstimated95% range

#1 in dimensionTop-10 median

IQ dimensions

Ranks are among models scored in each dimension. Ranges are model-based estimates, not calibrated confidence intervals.

DimensionIQPosition · 70–160RankEvidencevs Top 10

Abstract ReasoningAbstract

86estimated

#126 / 201

0/5

-40.5

Mathematical ReasoningMath

94estimated

#142 / 176

0/5

-53

Academic ReasoningAcademic

117measured

#116 / 203

5/6

-18.5

Programmatic ReasoningCoding

97measured

#150 / 177

2/6

-52.5

Computer UseComputer Use

101estimated

#118 / 176

0/5

-42

ReliabilityReliability

108measured

#119 / 177

6/6

-26

Knowledge WorkKnowledge Work

103measured

#106 / 176

4/6

-28.5

gemma-4-31b95% range (model-based, uncalibrated)Top-10 median#1 in dimension100 = average human

Benchmark results

17 of 39 Composite IQ benchmarks have a published result for gemma-4-31b. Benchmarks without a result stay listed so gaps are visible.

Abstract Reasoning IQ 86 · estimatedNo published results0/5

No published results yet for Abstract Reasoning. The 86 shown is estimated from gemma-4-31b's other dimensions and will be replaced as results come in. How estimates work

ARC-AGI-3ARC-AGI-2ARC-AGI-1Chess PuzzlesMystery Game Puzzles

Mathematical Reasoning IQ 94 · estimatedNo published results0/5

No published results yet for Mathematical Reasoning. The 94 shown is estimated from gemma-4-31b's other dimensions and will be replaced as results come in. How estimates work

FrontierMath Tier 4FrontierMath Tier 1-3ProofBench 1.1MathArenaAIME

Academic Reasoning IQ 117 · measuredBest: #77 of 147 on GPQA Diamond5/6

BenchmarkScorePercentileRankLeader · source

GPQA Diamond

Artificial Analysis v4.3.2 · Oct 27

85.7%

#77 / 147

Leader gpt-6-astra · 96.1%
Artificial Analysis v4.3.2 · Oct 27

SciCode

Artificial Analysis v4.3.2 · Oct 27

45.5%

#89 / 167

Leader opus-5.5 · 66.9%
Artificial Analysis v4.3.2 · Oct 27

Humanity's Last Exam

Artificial Analysis v4.3.2 · Oct 27

23.6%

#106 / 167

Leader opus-5.5 · 61.4%
Artificial Analysis v4.3.2 · Oct 27

CritPt

Artificial Analysis v4.3.2 · Oct 27

1.4%

#110 / 159

Leader Muse Wonder Full · 32.9%
Artificial Analysis v4.3.2 · Oct 27

MMMU-Pro

Artificial Analysis v4.3.2 · Oct 27

73.4%

#63 / 89

Leader fable-5.1 · 90.6%
Artificial Analysis v4.3.2 · Oct 27

Not yet reported

MMLU-Pro

Programmatic Reasoning IQ 97 · measuredBest: #76 of 104 on Terminal-Bench 4.02/6

BenchmarkScorePercentileRankLeader · source

Terminal-Bench 4.0

Artificial Analysis v4.3.2 · Oct 27

0.0%

#76 / 104

Leader Kimi K3.1 Pro · 67.2%
Artificial Analysis v4.3.2 · Oct 27

SWE-rebench

SWE-rebench · Sep 22

16.5%

#25 / 25

Leader fable-5 · 64.5%
SWE-rebench · Sep 22

Not yet reported

LiveCodeBenchIOI V2ProgramBenchMirrorCode

Computer Use IQ 101 · estimatedNo published results0/5

No published results yet for Computer Use. The 101 shown is estimated from gemma-4-31b's other dimensions and will be replaced as results come in. How estimates work

BrowseCompOSWorld-VerifiedMCP AtlasArena.ai Agent ArenaAgents' Last Exam

Reliability IQ 108 · measuredBest: #5 of 36 on FACTS Grounding6/6

BenchmarkScorePercentileRankLeader · source

FACTS Grounding

Full source leaderboard reconciliation · Oct 23

80.7%

#5 / 36

Leader gpt-6-sol · 85.2%
Full source leaderboard reconciliation · Oct 23

IFBench

Artificial Analysis v4.3.2 · Oct 27

75.6%

#23 / 96

Leader minimax-m3 · 82.9%
Artificial Analysis v4.3.2 · Oct 27

AA Long Context Reasoning v1.1

Artificial Analysis v4.3.2 · Oct 27

69.7%

#111 / 154

Leader kimi-k3 · 88.7%
Artificial Analysis v4.3.2 · Oct 27

BullshitBench v2

Full source leaderboard reconciliation · Oct 23

25.0%

#60 / 83

Leader opus-4.8 · 94.0%
Full source leaderboard reconciliation · Oct 23

AA Omniscience

Artificial Analysis v4.3.2 · Oct 27

-47.9

#145 / 158

Leader opus-5.5 · 46.4
Artificial Analysis v4.3.2 · Oct 27

SimpleQA Verified

Full source leaderboard reconciliation · Oct 23

10.4%

#80 / 80

Leader Gemini 4 Argon · 76.0%
Full source leaderboard reconciliation · Oct 23

Knowledge Work IQ 103 · measuredBest: #76 of 106 on GDP.pdf4/6

BenchmarkScorePercentileRankLeader · source

GDP.pdf

Artificial Analysis v4.3.2 · Oct 27

6.0%

#76 / 106

Leader Muse Wonder Full · 32.6%
Artificial Analysis v4.3.2 · Oct 27

AutomationBench-AA

Artificial Analysis v4.3.2 · Oct 27

4.9%

#77 / 106

Leader Gemini 4 Argon · 77.5%
Artificial Analysis v4.3.2 · Oct 27

GDPval-AA v2.1 normalized score

Artificial Analysis v4.3.2 · Oct 23

6.0%

#98 / 124

Leader opus-5.5 · 68.3%
Artificial Analysis v4.3.2 · Oct 23

AA-Briefcase v1.1 Elo

Artificial Analysis v4.3.2 · Oct 27

368

#94 / 111

Leader sonnet-5.5 · 1,812
Artificial Analysis v4.3.2 · Oct 27

Not yet reported

AA-AnalystAgentEnterpriseOps-Gym-AA

Release & availability

Release details on record for this model.

Released: Apr 2, 2026
Developer: Google
Weights: Open weightsDownloadable and self-hostable
Context: 256,000 tokens
API price: $0.12 in · $0.35 outper 1M tokens · full pricing

About this score

How much of gemma-4-31b's IQ is measured, and how the rest is estimated.

17/ 39benchmark results

4/ 7dimensions measured

3estimated

0/5Abstract

0/5Math

5/6Academic

2/6Coding

0/5Computer Use

6/6Reliability

4/6Knowledge Work

Bar height = share of benchmarks with resultsEstimatedNo score

Measured results are shown as published

Every score in the benchmark tables comes from its source unchanged. Composite IQ averages all seven dimensions equally.

Estimates fill the gaps

Abstract Reasoning, Mathematical Reasoning and Computer Use have no published results yet, so their IQ is estimated from gemma-4-31b's other dimensions.

A missing result is not a zero

Benchmark versions and evaluation harnesses can differ between sources, so small gaps between models may not be meaningful.

Read the scoring methodology →

Compare gemma-4-31b

Side-by-side pages for the models it is most often weighed against.

#1 overall

gpt-6.1-sol

IQ141 40 higherEffective cost$6.30 46.0× pricier

IQ101 same IQEffective cost$2.19 16.0× pricier

IQ94 7 lowerEffective cost$0.149 about the same

#2 overall

opus-5.5

IQ141 40 higherEffective cost$35.73 260.6× pricier

gemma-4-31bComparison model

Compare with any modelBrowse all comparisons →

Frequently asked questions

Quick answers about gemma-4-31b, drawn from the same data as the rest of this page.

How AI IQ scores models →Get this data from the API →

What is the IQ of gemma-4-31b?: gemma-4-31b has an estimated IQ of 101, which ranks #120 out of 176 public models on the AI IQ leaderboard. The score combines seven equally weighted reasoning dimensions using measured benchmark results and conservative estimates for missing coverage.
What is gemma-4-31b best at?: Among the dimensions with published benchmark results, gemma-4-31b ranks highest in Academic Reasoning (IQ 117, #116 of 203) and Knowledge Work (IQ 103, #106 of 176). Its lowest-ranked measured dimension is Programmatic Reasoning (IQ 97, #150 of 177).
What is the EQ of gemma-4-31b?: gemma-4-31b has an estimated EQ of 105. EQ measures emotional reasoning — how well a model reads and responds to emotional context — and is scored separately from IQ.
Which benchmarks has gemma-4-31b been tested on?: gemma-4-31b has published results on 17 of the 39 benchmarks behind AI IQ's Composite IQ, including FACTS Grounding, IFBench, GPQA Diamond, SciCode and Humanity's Last Exam.
How much does gemma-4-31b cost?: gemma-4-31b costs $0.12 per million input tokens and $0.35 per million output tokens. Adjusted for how many tokens it actually consumes on real tasks, its effective cost works out to $0.1371 per million input-plus-output tokens.
Is gemma-4-31b open source?: Yes — gemma-4-31b is an open-weights model: its trained weights are published, so it can be downloaded and self-hosted.
Who developed gemma-4-31b?: gemma-4-31b was developed by Google, based in the United States.
When was gemma-4-31b released?: gemma-4-31b was publicly released on April 2, 2026.
How large is the context window of gemma-4-31b?: gemma-4-31b supports a context window of 256,000 tokens — the amount of text it can consider in a single request.
How fast is gemma-4-31b?: In independent measurements, gemma-4-31b generates about 35 tokens per second, with a typical response time of about 64 seconds.