Google: Gemma 4 31B

google/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...

Prompt $/1M

$0.120

Completion $/1M

$0.370

Context

262,144 tok

Best provider

90 tok/s · SambaNova

Specs

Model id: google/gemma-4-31b-it
Display name: Google: Gemma 4 31B
Modality: text+image+video->text
Tokenizer: Gemma
Instruct type: —
Context length: 262,144
Max completion: 16,384
Prompt $/1M: $0.120
Completion $/1M: $0.370
Free tier: no

Capabilities

Function callingno
JSON modeno
System promptno
Visionyes
Moderatedno

Benchmark scores

Each axis is one benchmark, normalized to the best model in the field.

No benchmark data for this model.

Speed by provider

Provider	tok/s	TTFT	Total latency
SambaNovafastest	90	2,893 ms	3,487 ms
Artificial Analysis	67	825 ms	—
ModelRun	62	476 ms	666 ms
Friendli	56	5,236 ms	5,437 ms
WandB	36	397 ms	492 ms
Venice	30	1,134 ms	1,453 ms
Cerebras	28	393 ms	521 ms
SiliconFlow	21	2,452 ms	3,452 ms
Chutes	20	2,427 ms	5,461 ms
OpenInference	19	1,194 ms	1,820 ms
Parasail	17	1,775 ms	3,147 ms
Together	14	1,059 ms	1,553 ms
Phala	12	4,375 ms	7,118 ms
DeepInfra	10	888 ms	1,152 ms
Novita	10	994 ms	1,251 ms
Ambient	5	2,877 ms	4,510 ms
AkashML	4	4,466 ms	7,768 ms