Back to Models

Gemma 4 31B IT

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...

Parameters
N/A
Context
262,144 tokens
Released
Feb 4, 2026

Leaderboards

Average Score combining domain-specific Autobench scores; Higher is better

Performance vs. Industry Average

Intelligence

Gemma 4 31B IT is of lower intelligence compared to average (4.5), with an intelligence score of 4.3.

Price

Gemma 4 31B IT is cheaper compared to average ($1.13 per 1M Tokens) with a price of $0.02 per 1M Tokens.

Latency

Gemma 4 31B IT has a lower average latency compared to average (47.46s), with an average latency of 15.89s.

P99 Latency

Gemma 4 31B IT has a lower P99 latency compared to average (288.94s), taking 117.31s to receive the first token at P99 (TTFT).

Context Window

Gemma 4 31B IT has a smaller context window than average (453k tokens), with a context window of 262k tokens.

Gemma 4 31B IT - AutoBench