Ctrl K
Back to models

Google

Gemma 4 31B

  • Released
  • Open weights
  • Apache 2.0
Compare
Intelligence Index

Reasoning setting

14.7#60 of 83measured models
Measured · Artificial Analysis v4.3.2
API price from
$0.14 in / $0.40 out / 1M tokOpenRouter · Crusoe · BF16
Context window
256K tok

Performance profile

ReasoningRank #34 of 107
85.7%
GPQA Diamond
CodingRank #46 of 94
45.5%
SciCode

Capabilities

Context window

Modalities

Measured setting

Integration

Reasoning trade-offs

Source (opens in a new tab)
Intelligence Index v4.3.2
Gemma 4 31B
Cost per task (USD) · log scale
Gemma 4 31BReasoning
Intelligence Index
14.7
Cost / task
$0.00
Evaluation time
5.8m
Google API free tier

Pricing & access

API route
OpenRouter · Crusoe · BF16API
Tier & contextStandard

USD per 1M tokens

Input
$0.14
Output
$0.40
Cached input
$0.14

Run locally

GGUF
Context
Estimated memory

~26.3 GiB

RAM + VRAM

Estimate assumptions

One text chat at the selected context, with a 16-bit cache and app allowance. Operating system and image processing are extra; optimized caches can use less.

32 GiB RAMwithout GPU offload

Memory budgetOne text chat
Weights
17.4 GiB
Context cache
6.9 GiB16-bit cache
App allowance
2 GiB
Disk space
~27 GBIncludes image file
Available software
Ollama (opens in a new tab)LM Studio (opens in a new tab)Jan (opens in a new tab)KoboldCpp (opens in a new tab)llama.cpp (opens in a new tab)
Get GGUF (opens in a new tab)
Set up in Ollama · Q4_K_M · 8K context

Save the following as Modelfile. The context matches the estimate above.

FROM hf.co/lmstudio-community/gemma-4-31B-it-GGUF:Q4_K_M
PARAMETER num_ctx 8192

Then run these commands in your terminal:

ollama create llmcurrent-local -f Modelfile
ollama run llmcurrent-local

Ollama import instructions (opens in a new tab) · For LM Studio or Jan, open the app’s model search, enter lmstudio-community/gemma-4-31B-it-GGUF, choose Q4_K_M, and set context to 8,192 tokens in the app’s model settings. Software links above open each app’s setup guide.

Images need extra memoryvLLM setup (opens in a new tab)

Benchmarks

Full leaderboardSource (opens in a new tab)

Humanity’s Last Exam

Frontier reasoning

Gemma 4 31B · Reasoning#57 of 126

  1. 1Claude Opus 5.5Max61.4%
  2. 2Claude Fable 5.1Max59.1%
  3. 3Gemini 4 ArgonHigh57.1%
  5. 54Nemotron 3 Ultra 550B A55BReasoning28.4%
  6. 55Qwen3.5 122B A10BReasoning25.2%
  7. 56Nemotron 3.5 Bolt 74B-A9BReasoning24.7%
  8. 57Gemma 4 31BReasoning23.6%
  9. 58K2 Horizon MoVA 36B A4BReasoning23.4%
  10. 59Ling-3.0-flash-FinReasoning22.6%
  11. 60Qwen3.6 35B A3BReasoning22.2%

Peers use their best measured result on this benchmark.

Alternatives

Each model at its highest-scoring measured setting · Intelligence Index v4.3.2 and its evaluation cost per task.