head-to-head

MiniMax: MiniMax M3 vs Google: Gemma 4 26B A4B (free)

Side-by-side comparison of specs, pricing, benchmark scores, and task rankings. Updated 2026-11-07.

MiniMax: MiniMax M3 Google: Gemma 4 26B A4B (free)
Vendorminimaxgoogle
Quality Score100100
Benchmark Score78.048.4
Input Price$0.30/MFree
Output Price$1.20/MFree
Context Window1,048,576262,144
Max Output512,00032,768
Tool Calling✓✓
Structured Output✓✓
Reasoning Mode✓✓
Vision✓✓
Audio--
Benchmark Scores
ai_index73.342.4
ai_index_agentic58.318.1
ai_index_coding96.664.9
eqbench-56.3

The verdict

For a sample job of 1 million input tokens and 250,000 output tokens, Gemma 4 26B A4B (free) costs $0 and MiniMax M3 costs $0.60, so Gemma 4 26B A4B (free) is free for the same work.

MiniMax M3 reads up to 1,048,576 tokens at once, roughly 1,573 pages of text, against about 393 pages for Gemma 4 26B A4B (free).

On our blended benchmark score MiniMax M3 leads 78.0 to 48.4.

Across the 20 tasks where they differ, MiniMax M3 scores higher on 20 (SQL Generation, Code Review, Code Completion and others).

MiniMax M3 was released in May 2026 and Gemma 4 26B A4B (free) in April 2026.

Pick MiniMax M3 if you need long documents, benchmark performance or the stronger all-round task scores; pick Gemma 4 26B A4B (free) if you need lower cost.

Costs use current list prices per million tokens; page counts assume about 0.75 words per token and 500 words per page.

Who wins by task?

TaskMiniMax: MiniMax M3Google: Gemma 4 26B A4B (free)
SQL Generation 167 153
Code Review 162 150
Code Completion 133 131
Code Refactoring 160 148
Bug Fixing 173 156
Unit Test Generation 151 140
Code Documentation 142 136
Regex Writing 134 130
CI/CD Pipelines 141 133
Frontend Component Design 143 136
Data Analysis 164 148
CSV / Spreadsheet Cleanup 153 143
ETL Scripting 151 141
JSON Extraction 147 142
Bulk Data Labeling 135 133
OCR / Document Parsing 145 138
Table Extraction from PDFs 145 138
Long-Document Summarization 156 147
Short-Form Summarization 131 129
Blog Post Writing 138 132

Scores reflect capability match + benchmark data + pricing for each task. Methodology →

Related comparisons

Anthropic: Claude Haiku Latest vs Google: Gemma 4 26B A4B (free) Anthropic: Claude Opus 5 (batch) vs MiniMax: MiniMax M3 Anthropic: Claude Sonnet 5 (batch) vs MiniMax: MiniMax M3 Anthropic: Claude Sonnet 5 vs Google: Gemma 4 26B A4B (free) Anthropic: Claude Sonnet 5 vs MiniMax: MiniMax M3 Anthropic: Claude Sonnet Latest vs Google: Gemma 4 26B A4B (free) ByteDance Seed: Seed-2.0-Code vs MiniMax: MiniMax M3 ByteDance Seed: Seed 2.1 Turbo vs MiniMax: MiniMax M3

The Model Movers Report

One email every Friday, built from this site's own rankings: the current top five by benchmark score, every model released in the last seven days, and one note worked out from that week's numbers. You can unsubscribe from any issue with one click.