Model Terminal

gemma-4-31b-it-oQ8e-mtp (dynamicagency)

A community-hosted Hugging Face model artifact packaging Google DeepMind's Gemma 4 31B instruction-tuned weights with multi-token prediction (MTP) support for speculative decoding. The MTP assistant is a small, text-only drafter model used alongside the primary Gemma 4 31B IT model to accelerate inference throughput. Source: supabase.

Value score
Not scored
Context
tokens
Max output
tokens
Price

Capability radar

Identity

Openness
Modalities
Release
Knowledge cutoff
Deprecation
API docs

Benchmarks

No benchmark scores yet.

Compare nearby

Pricing

Input / 1M
Output / 1M
Speed