Language

Gemma 4 26B-A4B MTP Drafter (MoE)

Google's jointly-trained Multi-Token Prediction head for the Gemma 4 26B-A4B Mixture-of-Experts target. Pair via `--spec-type draft-mtp`; Unsloth measures ~1.15–1.2× speedup on MoE targets vs ~1.4–2.2× on dense.

Gemma LicenseSource verified

Specification

Model ID

gemma4-26b-a4b-mtp-draft

Family

gemma4

Modality

Language

Parameters

MTP head

Context

131,072 tokens

Quantization

BF16

Weights

1.1 GB

Minimum RAM

2 GB

Source

Weights and provenance.

Repository

unsloth/gemma-4-26B-A4B-it-GGUF

Registry license

Gemma License

Source license

apache-2.0

Access

Open — weights fetch without accepting additional terms.

The registry records this model as Gemma License while the source repository states apache-2.0. Both are shown here rather than one being preferred. Treat the stricter of the two as binding until the discrepancy is resolved upstream.

Serve it

# Pull the weights onto a node
tenzro model download gemma4-26b-a4b-mtp-draft

# Serve it
tenzro model serve gemma4-26b-a4b-mtp-draft

# Or query the registry entry over JSON-RPC
curl https://rpc.tenzro.xyz \
  -X POST -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,
       "method":"tenzro_getModel",
       "params":[{"model_id":"gemma4-26b-a4b-mtp-draft"}]}'