An unidentified free model called Ox Alpha appeared on OpenRouter and quickly drew massive usage from developers who couldn't determine its origin, sparking days of speculation about which lab built it. Z.ai revealed the model was actually its own GLM-5.3-Flash, running entirely on Chinese chips, and priced at 15 cents per million input tokens and 50 cents per million output tokens, with a launch discount cutting that in half through September 9.
venturebeat.com
· 2026-08-27
LM Studio has integrated Z.ai's newly released GLM-5.3-Flash model into its Bionic agent platform, adding multimodal image input and a 1 million-token context window. The company says the model outperforms its predecessor, GLM-5.2, while costing roughly 9-10 times less to run, and it's served from US-based servers with zero-data-retention enabled by default.
9to5mac.com
· 2026-08-27
GLM-5.3-Flash is a text-only model with a 400k token context window that scored 57 on the Artificial Analysis Intelligence Index, far above the comparable median of 18. It is priced at $0.15 per 1M input tokens and $0.50 per 1M output tokens, both below the median rates for similar models, with the full benchmark run costing $138.02.
artificialanalysis.ai
· 2026-08-26
Reddit users and Android Authority's own lux measurements indicate the Pixel 11 Pro's camera flash and flashlight are noticeably dimmer than the Pixel 10 Pro's, with one test showing 89 lux versus 123 lux. The likely culprit is a physical diffuser added around the camera visor to accommodate Google's new HiLight ambient LED array, which appears to soften the main flash's beam.
androidauthority.com
· 2026-08-25
At Hot Chips 2026, researchers Anurag Agarwal and Radhakrishna Giduthuri presented a tutorial on High Bandwidth Flash (HBF), a package-integrated flash memory technology built like HBM but functioning more like an SSD. Since no HBF hardware exists yet, the talk relied on simulations and projections, examining how software—especially inference frameworks like vLLM—would need to be redesigned to exploit HBF's large capacity despite its coarse-grained, storage-like access pattern.
chipsandcheese.com
· 2026-08-24
GoKawiil