all the models — AI benchmark observatory
← Models

Gemma 4 31B

Google · open weight · gemma-4-31b-it

Gemma 4 31B is Google DeepMind's flagship dense multimodal model with 31 billion parameters and a 256K context window. Ranks #3 among open models on Arena AI. Built from the same research as Gemini 3, it features Per-Layer Embeddings, Shared KV Cache, alternating sliding-window and global attention, and variable aspect ratio vision encoding. Achieves an estimated LMArena text score of 1452.

AIME 2026t2-benchMathVisionMMLU-ProGPQAMMMU-Pro

Benchmark scores

Pricing

  • DeepInfra$0.09 / $0.34
  • Novita$0.14 / $0.40
  • FriendliAI$0.14 / $0.40
  • Together$0.39 / $0.97

Input / output per 1M tokens

AA metrics

No Artificial Analysis link yet.

Arena Elo

  • No Arena snapshot linked.