cd /news/artificial-intelligence/qwen3-6-35b-a3b · home topics artificial-intelligence article
[ARTICLE · art-52724] src=tokenstead.ai ↗ pub= topic=artificial-intelligence verified=true sentiment=↑ positive

Qwen3.6 35B A3B

Alibaba releases Qwen3.6 35B A3B, a Mixture-of-Experts model with 35B total parameters and 3B active per token, featuring 256K native context and ranking #1 for agent workflows in 2026. The model achieves perfect tool-calling and zero catastrophic failures on a comprehensive agent benchmark, and is available under Apache 2.0 license.

read2 min views1 publishedJul 9, 2026
Qwen3.6 35B A3B
Image: Tokenstead (auto-discovered)

MoE enthusiast35B total, 3B active per token (MoE: 256 experts, 8 routed + 1 shared), hybrid Gated DeltaNet + Gated Attention layers. 256K native context, extensible to ~1M. The agentic flagship - ranked #1 local model for agent workflows in 2026 (perfect tool-calling, zero catastrophic failures on an 84-scenario / 16-category / 8-trial benchmark). Q4_K_M (~20GB) for GGUF runtimes; Q8_0 (~37GB) for full precision.

  • 35.0B
  • 262k
  • apache 2.0
  • Apr 2026

Scores #

Run it locally #

Per-quant memory needs and a static "can you run it?" reference - no rig entry required

Can you run it? - reference rigs

Rig Q4_K_M Q8_0
NVIDIA Jetson Orin NX 16GB

no -> cloudFit tiers use the same will-it-run logic as the rig finder. For comfortable fits, the badge reflects decode speed: fast >=20 t/s, ok 8-20 t/s, slow <8 t/s. t/s is a bandwidth estimate, not a measured benchmark.

How can a 24GB GPU run a 744B model? It does not load the model into VRAM. The quantized weights (e.g. ~410GB at Q4) sit in system RAM; the GPU only holds the small shared attention and router tensors and accelerates prompt processing. Because GLM 5.2 is a Mixture-of-Experts model, each token activates only ~40B of its 744B params, so llama.cpp streams just those active experts from system RAM to the GPU each token (the -cmoe

offload path).

That makes decode speed bound by system-RAM bandwidth, not GPU bandwidth - single digits on DDR4, which is why these rigs show 3-8 t/s even though they “fit.” A bigger GPU (e.g. 2x 3090) keeps more experts resident on-card and raises tok/s; a smaller GPU still runs it but pays the bandwidth tax. A 744B dense model could not run this way - only MoE’s small-active-params trick makes it possible.

Aggressive quants (1-2 bit) trade accuracy for size - roughly 17% accuracy loss at 2-bit vs full precision, and real long-context work often needs Q5 or Q6 even when lower quants “fit.”

Formula estimates here are conservative; real tuned setups can exceed them (one HN user reports ~6 tok/s on a 512GB DDR4 + 2x 3090 rig).

Download options #

Inference cost over time #

Data accumulates from the first daily sync - longer ranges populate over time. Prices come from OpenRouter snapshots, not a historical API.

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @alibaba 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/qwen3-6-35b-a3b] indexed:0 read:2min 2026-07-09 ·