Baseten slashes DeepSeek V4.1 Flash cached input 77% to $0.007 per 1M tokens
What happened
Between the Sep 21, 2026 and Sep 22, 2026 captures of Baseten's pricing page, Pulse recorded 2 changes: DeepSeek V4.1 Flash cached-input token price dropped from $0.03 to $0.007 per 1M tokens; input ($0.30) and output ($1.20) prices unchanged; Model API catalog reshuffled: GLM-5.3 Fast added; GLM 4.7, DeepSeek V4 Pro, Kimi K2.6, Kimi K2.7 Code, Inkling-Small, and Inkling removed from the per-token pricing table. Pulse files it as a pricing change (price decrease, feature change) with a downward direction. Pulse's latest capture (Oct 22, 2026) lists Basic at $1.32 DeepSeek V4 Pro 0813 input (per 1M tokens) + $0.3 Kimi K3 cache input (per 1M tokens).
It lands 48 days after Baseten's previous change, "Baseten: added Kimi K3, Inkling-Small & DeepSeek-V4-Flash-0731 model APIs" (Aug 5, 2026). Before that Pulse recorded "Baseten: Kimi K3 model added to Model APIs ($3.00/$0.30/$15.00 per 1M)" (Jul 28, 2026), "Baseten: GLM 5.2 cache-input price cut from $0.26 to $0.14/1M tokens" (Jul 27, 2026) and "Baseten: GLM-5.2 cache input price cut ($0.26→$0.14/1M tokens)" (Jul 25, 2026). Baseten has been tracked since 2026-02-26.
In the same window Zen Planner, Wodify, Pike13 and Mindbody also changed their AI & Machine Learning pricing pages. Those moves are listed further down, and the Pulse Agent can compare them side by side.
- — · Price−77%$0.03$0.007
- Pro · FeatureAddedNo GLM-5.3 FastGLM-5.3 Fast
| Plan | Field | Before | After | Change |
|---|---|---|---|---|
| — | Price | $0.03 | $0.007 | −77% |
| Pro | Feature | No GLM-5.3 Fast | GLM-5.3 Fast | Added |
Values as written on Baseten's public pricing page in the two captures Pulse compared; where the page states no number, the cells give the wording before and after.
Evidence
The two captures Pulse comparedFull-page captures of Baseten's pricing page (www.baseten.co) from Sep 21, 2026 and Sep 22, 2026; the change is marked around the middle of the page, where price reads $0.03 and then $0.007.
Sign in to see the evidence
The two full-page captures with this change marked up are free with an account.
Baseten, the last 90 days
Full history- Oct 17, 2026packagingBaseten adds Reflection Beam and Meta Muse Spark Open to Model APIs lineup
- Oct 16, 2026packagingBaseten adds Nemotron 3.5 Super VL at $0.20 per 1M input tokens
- Sep 28, 2026packagingBaseten adds GLM-5.3 Fast, retires 6 models, cuts DeepSeek V4.1 Flash cache price 77%
- Sep 22, 2026decreaseBaseten slashes DeepSeek V4.1 Flash cached input 77% to $0.007 per 1M tokens · this change
- Aug 5, 2026improvedBaseten: added Kimi K3, Inkling-Small & DeepSeek-V4-Flash-0731 model APIs
- Jul 28, 2026improvedBaseten: Kimi K3 model added to Model APIs ($3.00/$0.30/$15.00 per 1M)
- Jul 27, 2026decreaseBaseten: GLM 5.2 cache-input price cut from $0.26 to $0.14/1M tokens
- Jul 25, 2026decreaseBaseten: GLM-5.2 cache input price cut ($0.26→$0.14/1M tokens)
Across AI & Machine Learning in the same window
All AI & Machine Learning changes- Zen PlannerSep 21Zen Planner adds 3 bundled Packages plans, topping out at $348/mo Ultimate
- WodifyAug 4Wodify: Starting price increased $79→$99/month per location
- Pike13Jul 13Pike13: New FAQ discloses per-location billing for franchise networks
- MindbodyJul 13Mindbody Online: Starting price cut $99→$79/mo + AI Concierge now included on Ultimate
Questions about this change
Did Baseten raise prices in September 2026?
No — prices went down. Between Sep 21, 2026 and Sep 22, 2026 Baseten's pricing page showed 1 price cut: $0.03 → $0.007 (−77%).
What changed on Baseten's pricing page between Sep 21, 2026 and Sep 22, 2026?
Price: $0.03 → $0.007 (−77%); Pro feature: No GLM-5.3 Fast → GLM-5.3 Fast, added. Pulse compared the captures it took on Sep 21, 2026 and Sep 22, 2026; the before/after table above lists each item as written on the page.
How often does Baseten change its pricing?
Pulse recorded 4 other pricing-page changes for Baseten in the 90 days before this one; the previous change was "Baseten: added Kimi K3, Inkling-Small & DeepSeek-V4-Flash-0731 model APIs" on Aug 5, 2026; Baseten has been tracked since 2026-02-26.