Qwen3 Coder Next qwen3-coder-next

qwen · qwen coding

unranked — no quality signals yet

params
79.7B
arch
moe
context
256k
license
apache-2.0 open weights
released
Jan 2026
downloads/30d
434.1k

# architecture

attn shared GQA 16q/2kv router top-10 of 512 expert ×512 expert out ×48 layers ctx 262,144 79.7B pool
layers
48
d_model
2048
heads
16q / 2kv
head dim
256
experts
top-10 of 512
vocab
151936
family
qwen3_next

signal path of one layer, generated from the registry's structured fields — dimension lines quote the model's real numbers; an MoE trace forks at the router, a dense trace runs straight through. Geometry fields come from the repo's config.json.

# trend

OGM score · last 112 days
OGM score over 112 days
overall rank (up is better)
Overall rank over 112 days

methodology changed during this history window; score movement across that boundary is not model movement. See methodology v6.

# where to run

providerquantctx$/M in$/M out$/M cacheprice srctpsuptime✓
Aihubmix 256k $0.14$0.55— via litellm — —
Parasailbf16 256k $0.12$0.80$0.07 via openrouter — 99.9%
Novita 256k $0.20$1.50— via litellm — —
Amazon Bedrock — $0.50$1.20— bedrock — — ✓
Together 256k $0.50$1.20— via modelsdev — —

sorted by blended price ((3·input + output) / 4 per 1M) · ✓ = the provider's own catalog confirms the offer · "via …" prices are what the aggregator routing the offer charges, not the provider's own list price

source aliases
aa
qwen3-coder-next
epoch
Qwen3-Coder-Next
hf
unsloth/Qwen3-Coder-Next-GGUF
openrouter
qwen/qwen3-coder-next