The 30b model now fits in a phone. mediatek's dimensity 9600 pro, out september 15, is the first…
the 30b model now fits in a phone. mediatek's dimensity 9600 pro, out september 15, is the first 2nm smartphone soc, with an npu that lifts llm prefill throughput 51% and runs mixture-of-experts models up to 30b on-device.
the data center moved into your pocket.
Context
MediaTek's press release of 15 September 2026 announces the Dimensity 9600 Pro on a 2nm process, saying it is the first company to announce reaching that threshold. It says the NPU 1090 delivers 51 percent higher LLM prefill performance and 55 percent higher token generation per watt and supports on-device applications including models up to 30B parameters, with LPDDR6 memory and UFS 5.0 for the first time.
The announcement and the 51 percent are first-party and vendor-reported, higher than an unstated baseline, and the footnote was not read. First 2nm is MediaTek's own claim about announcing, and other vendors were not checked; an announcement is not device availability. Mixture-of-experts does not appear in the text read, which says models up to 30B parameters. No quantization, memory use, speed, named model or demo is given, so fits in a phone is not shown. The data center moved into your pocket is the author's take.
Related work
- Earlier note on a ternary 27B model and phones ↗A different on-device claim, with measured laptop and GPU speeds.
Watch next
- MediaTek's footnote and benchmark method, first devices, and a demo of a named 30B model with memory and speed.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 21 September 2026 at 15:17 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →