Qwen 35B Successfully Runs on a Smartphone
Severe_Post_2751 · reddit · 2026-07-18
The author tested a proprietary Qwen 35B MoE runtime on a Samsung S26 Ultra, claiming that without sacrificing accuracy, the activated model footprint fits within the device's memory constraints.
Early results show that after optimization, input processing is about 90 tokens/s, while output generation is around 8 tokens/s. The author emphasizes they are self-taught in AI/ML without a formal PhD, mentions having compute resources to continue testing, and invites collaboration. They also noted having submitted 4 arXiv papers, which are still under review due to a lack of institutional affiliation.
More from Infra
- Anthropic S-1 reveals 48 cents of compute cost per dollar of Q2 revenue — nmasc_ · 2026-10-26
- Persistent AI joins UR network to route private AI inference with encrypted, IP-hidden traffic — 0xSammy · 2026-10-26
- Anthropic's disclosed compute commitments: $114.5B Google, $161.2B Broadcom top the list — BenBajarin · 2026-10-26
- SpaceX strikes $40B deal with Apollo Global to lease AI chips and data center hardware — Polymarket · 2026-10-26
- SpaceX AI revenue up 830%, Claude designs 318K antivenom proteins in 64 hours — PeterDiamandis · 2026-10-26
- Meta plans up to $95 billion stock sale to fund AI buildout — The Decoder · 2026-10-26