Google DeepMind

Gemma 4 31B

Type
language model
Context
262K tokens
Max output
8K
Released
2 April 2026
API string
google/gemma-4-31b
Weights
open

Prices by provider

Provider Input, $ per 1M Output, $ per 1M In our data since
bedrock_mantle $0.14 $0.4 31 Jul 2026
neuralwatt $0.144 $0.42 31 Jul 2026
Cerebras $0.99 $1.49 31 Jul 2026

Only the base tier and only per-token prices are shown. Batch, discounted and cached rates, as well as prices per image or per second of video, do not go into this table: they cannot stand in the same column as a price per million tokens.

The date in the last column is the day this price first entered our collection. The price may well be older: before that day we simply were not recording it. It has not changed since — otherwise a new row with a new date would stand in its place.

Measurement results

Task set Result Run conditions Measured by
Arena Score, mathematics 1,466.07 1,438–1,494 — —
Arena Score in Russian 1,461.39 1,439–1,484 — —
Arena Score in Spanish 1,457.33 1,419–1,496 — —
Arena Score, programming 1,454.93 1,440–1,470 — —
Arena Score, expert questions 1,451.71 1,425–1,478 — —
Arena Score, multi-turn dialogue 1,450.83 1,433–1,469 — —
Arena Score in English 1,448.17 1,437–1,459 — —
Arena Score, hard prompts 1,445.76 1,436–1,456 — —
Arena Score, long queries 1,443.42 1,429–1,457 — —
Arena Score, overall 1,441.69 1,434–1,449 — —
Arena Score in French 1,437.84 1,400–1,475 — —
Arena Score, instruction following 1,432.14 1,418–1,446 — —
Arena Score, creative writing 1,417.17 1,398–1,436 — —
Arena Score, web development 1,363.39 1,355–1,371 — —
Arena Score, understanding diagrams 1,281 1,273–1,289 — —
Arena Score, text recognition in images 1,278.72 1,272–1,285 — —
Arena Score, working with images 1,271.6 1,265–1,278 — —
Arena Score, agent tasks -0.16 0–0 — —