Gemma 4 31B IT
google/gemma-4-31b-it- Family · Gemma 4
- Open weights
- Image · Text · Video → Text
- Tools · JSON · Vision
Spec & list price
Context256K
Max output32K
Input / 1M$0.20
Output / 1M$0.402× input
Usage
Tokens / day67.5B▼ 12% 7d
Usage rank#41
Requests / day8.2M
Tool calls / req0.02
Per request · 30 days
Avg input3.0K0.2× median
Avg output2680.4× median
Cache hit29%0.4× median
Reasoning share13%0.2× median
Tool-call failures1.9%about median
Consumption
Peak 92.0B 20 OctLow 44.9B 31 OctRank today #41
News & trends
1 mentions this week ▼ 50%this model · videos, posts, articles · 30 days
Trending · 14 days · Gemma 4 family
- Gemma-4-31B-Humanlike-Chat: 2.0's improved human voice on official Gemma 4 31B314 upvotes
- Qwen 3.8 27B vs Gemma 4 31B on RTX 5090: 4 Real Coding Tests1.8K views
- Qwen 3.8 27B vs Gemma 4 31B on RTX 5090: 4 Real Coding Tests6.2K views
- Gemini 3.6 Flash won a nuclear last stand without breaking a single pact. Google's open Gemma 4 31B came 2nd of 43.68 upvotes
News · latest from Google
- Gemini 4 Carbon is the new #1 on the Artificial Analysis Intelligence Index v529 Oct
- Gemini 4 Carbon (High) Scores 59 on Artificial Analysis v528 Oct
- Gemini 4 Carbonblog.google · 28 Oct
- Gemini 4 Carbon: Our next frontier modelGoogle DeepMind · 28 Oct
Benchmarks
Benchmark data: Artificial AnalysisOutput speed#287/33332 tok/s2.1× slower than median
Fastest provider107 tok/sbest of 17 providers
First token#1390.84 s1.2× quicker than median
Family progress
Intelligence Index
- Family releases14.2
Technical details
- Tokenizer
- Gemma
- License
- Apache 2.0
- Parameters
- 31.3B
- Moderation
- none
- Downloads · likes
- 8.3M · 4.3K
Request parameters 19
- tools
- tool_choice
- response_format
- structured_outputs
- reasoning
- include_reasoning
- temperature
- top_p
- top_k
- min_p
+9 more
- max_tokens
- stop
- seed
- frequency_penalty
- logit_bias
- logprobs
- presence_penalty
- repetition_penalty
- top_logprobs
When to use
Best for
- Self-hosted chat with image and video input
- Private deployments with tool calling
- Fine-tuning a capable open model
- Structured data extraction from images
Think twice for
- Very large outputs (short output cap)
- Hardest frontier reasoning tasks
- Audio input
Research papers
- Discovering Jailbreaks via Curriculum of Models28 Oct
- Evolving Self-tuning Benchmarks for Long-Horizon Tool Use19 Oct
- TxnQA: Benchmarking Analytical Reasoning and Evidence Grounding over Transaction Event Sequences19 Oct
- Screening LLM Judges Without Ground Truth: A Preliminary Evaluation of an AI Dental Hygiene Coach17 Oct
- HeRE: A Novel Benchmark for Hebrew Relation Extraction via LLM-Guided Denoising of Knowledge Graph Alignments15 Oct
- Calibrated Evidence Inference With a Small, Permissively Licensed Open Model4 Oct
- Asking Earns Nothing: Scoring the Decision to Act in BFCL Multi-Turn3 Oct
- ATTRICITE: Training an Open 4B Model for Citation Recovery toward Faithful Attribution13 Sep
Same family
Gemma 4- 8.9
- 14.2
- 16.7
- 14.7
- 7.8
Using Gemma 4 31B IT at volume?
Sign in to ask for a volume discount. You will continue in the OSMEXO dashboard.