The price war has an exception. zhipu's glm-5.3-flashx, out september 18, runs 200 tokens per…
the price war has an exception. zhipu's glm-5.3-flashx, out september 18, runs 200 tokens per second, five times faster than the prior version, and costs 2.5 times more, sending zhipu's stock up 5.34% the same day.
speed still commands a premium.
Context
PANews (18 September 2026), citing Zhipu's official WeChat account, reports GLM-5.3-FlashX launched with the API open and a maximum inference speed of 200 tokens per second on 100,000 domestic chips; the base GLM-5.3-Flash was open-sourced on 26 August. OpenRouter's page lists FlashX as released 18 September 2026 at 0.37 dollars input, 1.25 output and 0.09 cache read per million tokens, and its live throughput reading for the Z.ai row was 78 tokens per second at fetch time. Vercel's AI Gateway page lists 0.37 input and 1.25 output, and a secondary page of 24 September gives the same prices. Lookonchain (18 September) reports Zhipu up over 7 percent in afternoon trading citing Bitget data.
The 0.37 and 1.25 prices are consistent across router and secondary pages, and they are about 2.5 times GLM-5.3-Flash's list prices of 0.15 input and 0.50 output, so costs 2.5 times more fits when read as a new tier against Flash's list price; Flash was on a 50 percent promotion until 9 September per Z.ai's page. They are not shown as a rise in a prior FlashX price. Z.ai's own pricing page as fetched on 4 October has no GLM-5.3-FlashX row, so the price is not confirmed on the first-party page. The 200 tokens per second is a vendor-stated maximum, and the router's 78 is a different window and basis, so it does not confirm the peak or the 5x comparator, which is vendor-reported with no stated baseline. The 5.34 percent close was not found, since the source says over 7 percent in the afternoon, and a same-day move is not shown to be caused by the launch. Speed still commands a premium is the author's take.
Related work
- Earlier note on the FlashX price ↗Same model and price claim; the sources read describe a new tier.
- Earlier note on GLM-5.3-FlashX as a speed tier ↗Same model.
Watch next
- Zhipu's first-party FlashX announcement and rate card, and the 18 September closing price.
Sources
- Zhipu launches GLM-5.3-FlashX (PANews, 18 Sep 2026)panews.io
- Zhipu stock move (Lookonchain, 18 Sep 2026)lookonchain.com
- Z.ai pricingdocs.z.ai
- GLM-5.3-FlashX (OpenRouter)openrouter.ai
- GLM-5.3-FlashX (Vercel AI Gateway)vercel.com
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 21 September 2026 at 14:25 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 →