Google: Gemma 4 31B

Image input Text input Video input Text output Free Option
Author's Description

Gemma 4 31B Instruct is Google DeepMind's 30.7B dense multimodal model supporting text and image input with text output. Features a 256K token context window, configurable thinking/reasoning mode, native function...

Key Specifications
Cost
$$
Context
262K
Parameters
31B
Released
Apr 02, 2026
Speed
★
Ability
★★★★★
Reliability
★★★
Supported Parameters

This model supports the following parameters:

Frequency Penalty Presence Penalty Top P Max Tokens Seed Tools Include Reasoning Temperature Stop Logprobs Top Logprobs Response Format Tool Choice Structured Outputs Reasoning
Features

This model supports the following features:

Reasoning Tools Structured Outputs Response Format
Performance Summary

Gemma 4 31B Instruct, Google DeepMind's 30.7B dense multimodal model, demonstrates strong overall performance, particularly in reliability and cost-effectiveness. While its speed ranking indicates longer response times, placing it in the 15th percentile, it consistently provides cost-effective solutions, ranking in the 75th percentile for price. The model exhibits exceptional reliability with a 99% success rate across benchmarks, indicating minimal technical failures. Analysis of benchmark results reveals key strengths in hallucination prevention and ethical reasoning, achieving perfect 100% accuracy in both categories. It also excels in instruction following (88.7% accuracy, 95th percentile) and coding (93.9% accuracy, 85th percentile), aligning with its description as strong in these areas. General knowledge (98.7%) and email classification (98.9%) also show robust performance. While reasoning (82.2%) and mathematics (93.2%) are solid, they represent areas with slightly more room for improvement compared to its top-tier performances. The model's multimodal capabilities and extensive context window (256K tokens) further enhance its versatility.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.14
Completion $0.4

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Novita
Novita | google/gemma-4-31b-it-20260402 262K $0.14 / 1M tokens $0.4 / 1M tokens
Parasail
Parasail | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
AkashML
AkashML | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
Venice
Venice | google/gemma-4-31b-it-20260402 256K $0.2 / 1M tokens $0.4 / 1M tokens
AkashML
AkashML | google/gemma-4-31b-it-20260402 131K $0.2 / 1M tokens $0.4 / 1M tokens
Together
Together | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
DeepInfra
DeepInfra | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
DeepInfra
DeepInfra | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
Chutes
Chutes | google/gemma-4-31b-it-20260402 131K $0.12 / 1M tokens $0.37 / 1M tokens
Friendli
Friendli | google/gemma-4-31b-it-20260402 262K $0.14 / 1M tokens $0.4 / 1M tokens
DeepInfra
DeepInfra | google/gemma-4-31b-it-20260402 262K $0.135 / 1M tokens $0.51 / 1M tokens
SiliconFlow
SiliconFlow | google/gemma-4-31b-it-20260402 262K $0.75 / 1M tokens $1 / 1M tokens
Ambient
Ambient | google/gemma-4-31b-it-20260402 65K $0.2 / 1M tokens $0.4 / 1M tokens
Phala
Phala | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
Parasail
Parasail | google/gemma-4-31b-it-20260402 262K $0.15 / 1M tokens $0.4 / 1M tokens
WandB
WandB | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
SambaNova
SambaNova | google/gemma-4-31b-it-20260402 131K $0.38 / 1M tokens $1.15 / 1M tokens
ModelRun
ModelRun | google/gemma-4-31b-it-20260402 262K $0.75 / 1M tokens $1 / 1M tokens
Cerebras
Cerebras | google/gemma-4-31b-it-20260402 131K $0.2 / 1M tokens $0.4 / 1M tokens
OpenInference
OpenInference | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
Morph
Morph | google/gemma-4-31b-it-20260402 175K $0.2 / 1M tokens $0.4 / 1M tokens
CoreWeave
CoreWeave | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
WandB
WandB | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
Crusoe
Crusoe | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
NextBit
NextBit | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
DeepInfra
DeepInfra | google/gemma-4-31b-it-20260402 131K $0.2 / 1M tokens $0.4 / 1M tokens
CoreWeave
CoreWeave | google/gemma-4-31b-it-20260402 262K $0.1 / 1M tokens $0.34 / 1M tokens
Crusoe
Crusoe | google/gemma-4-31b-it-20260402 262K $0.14 / 1M tokens $0.4 / 1M tokens
Venice
Venice | google/gemma-4-31b-it-20260402 256K $0.12 / 1M tokens $0.36 / 1M tokens
Io Net
Io Net | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
Reka
Reka | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
DekaLLM
DekaLLM | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
DeepInfra
DeepInfra | google/gemma-4-31b-it-20260402 262K $0.16 / 1M tokens $0.32 / 1M tokens
DeepInfra
DeepInfra | google/gemma-4-31b-it-20260402 262K $0.3 / 1M tokens $0.6 / 1M tokens
PrimeIntellect
PrimeIntellect | google/gemma-4-31b-it-20260402 262K $0.174 / 1M tokens $0.292 / 1M tokens
NextBit
NextBit | google/gemma-4-31b-it-20260402 262K $0.15 / 1M tokens $0.4 / 1M tokens
Modal
Modal | google/gemma-4-31b-it-20260402 131K $0.34 / 1M tokens $0.1 / 1M tokens
PrimeIntellect
PrimeIntellect | google/gemma-4-31b-it-20260402 262K $0.2 / 1M tokens $0.4 / 1M tokens
KT
KT | google/gemma-4-31b-it-20260402 262K $0.1 / 1M tokens $0.34 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by google