VerdictPal
Join

Models / Google

Gemma 4 12B (Reasoning)

Gemma · Released 2026-06-07

Google's smaller Gemma 4 reasoning variant, evaluated by Artificial Analysis on 7 Jun. Treat it as a compact open-weights reasoning candidate until the local benchmark ledger has sourced rows.

Small open-weights reasoning model for local experiments, classroom demos, and routing tests before paying for frontier APIs.

#38 of 51 on AA Index · current snapshot
22 AA Index · 10 benchmark rows · 1 sourcesPosition is benchmark-specific — not a cross-family or cross-source ranking.
Composite42
Context128K
Input / 1M—
Output / 1M—
Knowledge cutoffopen
Output speed129 t/s
TTFT (default API)1.46s
Statuslightweight
ReasoningCodingAgenticLong context
Reasoning
75
Coding
38
Agentic
36
Long context
55

51average across 4 tested families

Benchmark placements

Where Gemma 4 12B (Reasoning) places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.

BenchmarkFamilyRankScoreSourceDate
AA output speedPerformance#19/ 69Artificial Analysis2026-07-21
τ³-Bench BankingAgentic#19/ 6436Artificial Analysis2026-07-21
IFBenchReasoning#23/ 5774Artificial Analysis2026-07-21
AA time to first tokenPerformance#36/ 69Artificial Analysis2026-07-21
AA-LCRLong context#52/ 6455Artificial Analysis2026-07-21
Humanity's Last ExamReasoning#52/ 6515Artificial Analysis2026-07-21
SciCodeCoding#53/ 6538Artificial Analysis2026-07-21
Artificial Analysis Intelligence IndexReasoning#54/ 6722Artificial Analysis2026-07-21
GPQA DiamondReasoning#57/ 6975Artificial Analysis2026-07-21
Terminal-BenchAgentic#61/ 6918Artificial Analysis2026-07-21
Family context

The three highest-scoring carded models in each capability family. Where Gemma 4 12B (Reasoning) shows up, it's highlighted.

ReasoningGemma 4 12B (Reasoning) · 75
  1. 1
    Claude Mythos Preview
    95
  2. 2
    GPT-5.6 Sol
    94
  3. 3
    Kimi K3
    94
CodingGemma 4 12B (Reasoning) · 38
  1. 1
    GPT 5.5
    96
  2. 2
    Claude Opus 4.8
    96
  3. 3
    Claude Opus 4.6
    96
AgenticGemma 4 12B (Reasoning) · 36
  1. 1
    GPT 5.5
    100
  2. 2
    Claude Opus 4.8
    100
  3. 3
    Claude Fable 5
    100
Long contextGemma 4 12B (Reasoning) · 55
  1. 1
    Kimi K3
    75
  2. 2
    GPT-5.1
    75
  3. 3
    GPT-5.6 Sol
    74
Newest receipts

Benchmark rows added to the public ledger for Gemma 4 12B (Reasoning) in the last 120 days. Older rows live in the full table below.

10 recent
BenchmarkFamilyScoreSourceDays ago
AA-LCRLong context55.33%Artificial Analysis2
AA output speedPerformance129 t/sArtificial Analysis2
AA time to first tokenPerformance1.46sArtificial Analysis2
Artificial Analysis Intelligence IndexReasoning22%Artificial Analysis2
GPQA DiamondReasoning75.3%Artificial Analysis2
Humanity's Last ExamReasoning14.8%Artificial Analysis2
IFBenchReasoning73.54%Artificial Analysis2
SciCodeCoding38.2%Artificial Analysis2
τ³-Bench BankingAgentic36.26%Artificial Analysis2
Terminal-BenchAgentic18.18%Artificial Analysis2
Family coverage

A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.

Reasoning75
Coding38
Agentic36
Long context55
Price vs. performance
0255075100$0.10$0.30$1$3$10$30input USD / 1M tokens · log scale →← intelligence

Hover or tab any dot for name, score, and input price.

AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMiniMaxxAIMetaZhipuXiaomiMeituanCohere

Full benchmark ledger

Every catalog benchmark for Gemma 4 12B (Reasoning). Scores link to the original source; gaps mean no public row exists yet.

10 sourced rows

10 of 32 catalog benchmarks have a sourced row for Gemma 4 12B (Reasoning).

Reasoning

Coding

Agentic

Long context

Performance

Sources