Language

Qwen 3.8 2.4T-A95B (MoE)

Qwen 3.8 flagship MoE — 2.4T total, 95B active per token, thinking-only. 256K native context extensible to ~1M. Multi-node MoE-sharded serving only (10-shard UD-Q1_0 GGUF; first shard is the load entry). Built-in single MTP layer for self-speculative decoding.

qwen3.8-maxSource verified

Specification

Model ID

qwen3.8-2.4t-a95b

Family

qwen3.8

Modality

Language

Parameters

2.4T (MoE, 95B active)

Context

262,144 tokens

Quantization

UD-Q1_0

Weights

369.7 GB

Minimum RAM

450 GB

Source

Weights and provenance.

Repository

unsloth/Qwen3.8-2.4T-A95B-GGUF

Registry license

qwen3.8-max

Source license

other

Access

Open — weights fetch without accepting additional terms.

The registry records this model as qwen3.8-max while the source repository states other. 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 qwen3.8-2.4t-a95b

# Serve it
tenzro model serve qwen3.8-2.4t-a95b

# 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":"qwen3.8-2.4t-a95b"}]}'

Same family

Qwen 3.8 27B

27B (dense)

Qwen 3.8 27B DFlash-2 Drafter