RUNLOCALAIv38
→WILL IT RUNBEST GPUCOMPARETROUBLESHOOTSTARTPULSEMODELSHARDWARETOOLSBENCH
RUNLOCALAI

Operator-grade instrument for local-AI hardware intelligence. Hand-written verdicts. Real benchmarks. Reproducible commands.

OP·Fredoline Eruo
DIR
  • Models
  • Hardware
  • Tools
  • Benchmarks
  • Will it run?
GUIDES
  • Best GPU
  • Best laptop
  • Best Mac
  • Best used GPU
  • Best budget GPU
  • Best GPU for Ollama
  • Best GPU for SD
  • AI PC build $2K
  • CUDA vs ROCm
  • 16 vs 24 GB
  • Compare hardware
  • Custom compare
REF
  • Systems
  • Ecosystem maps
  • Pillar guides
  • Methodology
  • Glossary
  • Errors KB
  • Troubleshooting
  • Resources
  • Public API
EDITOR
  • About
  • About the author
  • Changelog
  • Latest
  • Updates
  • Submit benchmark
  • Send feedback
  • Trust
  • Editorial policy
  • How we make money
  • Contact
LEGAL
  • Privacy
  • Terms
  • Sitemap
MAIL · MONTHLY DIGEST
Get monthly local AI changes
Monthly recap. No spam.
DISCLOSURE

Some links on this site are affiliate links (Amazon Associates and other first-class retailers). When you buy through them, we earn a small commission at no extra cost to you. Affiliate links do not influence our verdicts — there are cards we rate highly that we don't have affiliate relationships with, and cards that sell well that we refuse to recommend. Read more →

SYS · ONLINEUPTIME · 100%2026 · operator-owned
RUNLOCALAI · v38
← /pulse/qwen-3-235b-a22b-permissive-license-confirmed
ADVISORYINDUSTRY MOVE·2026-04-29

Qwen 3 235B-A22B confirmed under Apache 2.0 — commercial deployment unlocked

▼ WHAT HAPPENED

Alibaba confirmed Qwen 3 235B-A22B (the 235B-total / 22B-active MoE flagship) ships under Apache 2.0 — fully permissive commercial use. This makes it the largest open-weight model with truly permissive licensing, surpassing DeepSeek V3 (which has additional terms in the DeepSeek License) and Llama 3.3 70B (Llama Community License with usage restrictions). For commercial production deployments without licensing review overhead, this is meaningful.

▼ OPERATOR ANGLE

**For new commercial production**: Qwen 3 235B-A22B is now the strongest "no license worries" pick at the 100B+ tier. MoE serving cost (22B active) competitive with 70B dense models. **For replacing Llama 3.3 70B in commercial pipelines**: real upgrade path on capability with cleaner licensing. Migration cost is the serving framework switch (vLLM MoE routing) and prompt re-tuning. **For evaluation**: run your specific workload — Qwen 3 reasoning is comparable to DeepSeek V3 at lower serving cost; English-chat polish is on par with Llama 3.3. See [Qwen 3 235B-A22B verdict](/models/qwen-3-235b-a22b), [Qwen family hub](/families/qwen), [Llama family hub](/families/llama) for license comparison.
SOURCE: https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct[HF-MODEL]

▼ ENTITIES REFERENCED

MODELQwen 3 235B-A22BMODELDeepSeek V3 (671B MoE)TASKReasoning & MathTASKText GenerationFAMILYDeepSeekFAMILYLlamaFAMILYQwen
[pulse item] · runlocalai.co/pulse/qwen-3-235b-a22b-permissive-license-confirmed