Gemma 4 31B (Reasoning)
GoogleGemmaOpen WeightApache 2.0 · Commercial OK
Description
Gemma 4 31B is Google DeepMind's flagship dense multimodal model with 31 billion parameters and a 256K context window. Ranks #3 among open models on Arena AI. Built from the same research as Gemini 3, it features Per-Layer Embeddings, Shared KV Cache, alternating sliding-window and global attention, and variable aspect ratio vision encoding. Achieves an estimated LMArena text score of 1452.
Release Date
2026-04-02
Parameters
30.7B
Context Length
262K
Modalities
image, text, video
Capability Radar
36
general
39
coding
86
reasoning
58
scienceest.
90
agents
70
multimodal
Science uses a reasoning proxy when dedicated science benchmarks are unavailable.
Rankings
| Domain | #Rank | Score | Source |
|---|---|---|---|
| Agents & Tools | 1 | 86.0 | LS |
| Code Ranking | 63 | 68.0 | AA |
| General Ranking | 90 | 69.0 | AA |
| Science | 51 | 71.0 | AA |
Benchmark Scores (LLM Stats)
Agents
t2-bench
86.4%SR
Biology
GPQA
84.3%SR
Finance
MMLU-Pro
85.2%SR
General
MMMLU
88.4%SR
LiveCodeBench v6
80.0%SR
MMMU-Pro
76.9%SR
BIG-Bench Extra Hard
74.4%SR
MRCR v2
66.4%SR
Healthcare
MedXpertQA
61.3%SR
Math
AIME 2026
89.2%SR
MathVision
85.6%SR
Humanity's Last Exam
26.5%SR
AA Evaluation Indices
Intelligence Index39.2
Coding Index38.7
Gpqa0.9
Ifbench0.8
Lcr0.6
Tau20.6
Scicode0.4
Terminalbench Hard0.4
Hle0.2
LLM Stats Category Scores
Tool Calling90
Agents90
Finance90
Legal90
Biology80
Chemistry80
General80
Language80
Physics80
Math70
Multimodal70
Reasoning70
Vision60
Healthcare60
Long Context60
Pricing
Input PriceFree
Output PriceFree
Blended Price (3:1)Free
Speed
Tokens/sec34.9 tokens/s
Time to First Token1.02s
Time to Answer58.33s
Available Providers
(LS internal units)| Provider | Input Price | Output Price |
|---|---|---|
| Novita | 140K | 400K |