head-to-head

Google: Gemma 4 31B (free) vs Google: Gemma 4 31B

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

Google: Gemma 4 31B (free) Google: Gemma 4 31B
Vendorgooglegoogle
Quality Score100100
Benchmark Score53.753.7
Input PriceFree$0.20/M
Output PriceFree$0.40/M
Context Window262,144262,144
Max Output32,76816,384
Tool Calling✓✓
Structured Output✓✓
Reasoning Mode✓✓
Vision✓✓
Audio--
Benchmark Scores
ai_index48.448.4
ai_index_agentic23.823.8
ai_index_coding71.771.7
eqbench57.557.5

The verdict

For a sample job of 1 million input tokens and 250,000 output tokens, Gemma 4 31B (free) costs $0 and Gemma 4 31B costs $0.30, so Gemma 4 31B (free) is free for the same work.

Both read up to 262,144 tokens at once, roughly 393 pages of text.

Across the 8 tasks where they differ, Gemma 4 31B (free) scores higher on 8 (SQL Generation, Code Completion, Code Documentation and others).

Pick Gemma 4 31B (free) if you need lower cost or the stronger all-round task scores.

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?

TaskGoogle: Gemma 4 31B (free)Google: Gemma 4 31B
SQL Generation 156 156
Code Completion 131 131
Code Documentation 137 137
Regex Writing 131 131
JSON Extraction 143 143
Bulk Data Labeling 133 133
Short-Form Summarization 129 129
Blog Post Writing 133 133

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

Related comparisons

Anthropic: Claude Haiku Latest vs Google: Gemma 4 31B Anthropic: Claude Haiku Latest vs Google: Gemma 4 31B (free) Anthropic: Claude Sonnet 5 vs Google: Gemma 4 31B Anthropic: Claude Sonnet 5 vs Google: Gemma 4 31B (free) Anthropic: Claude Sonnet Latest vs Google: Gemma 4 31B Anthropic: Claude Sonnet Latest vs Google: Gemma 4 31B (free) Google: Gemini 3.1 Flash Lite vs Google: Gemma 4 31B Google: Gemini 3.1 Flash Lite vs Google: Gemma 4 31B (free)

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.