DeepSeek cuts prices, pauses fundraising
A cheap new model ships as reports surface of a stalled raise.
What's happening
DeepSeek released DeepSeek-V4-Flash (0731), which Chinese coverage says hits roughly a 98% cache-hit rate for cost efficiency, and Artificial Analysis scored it 50 on its Intelligence Index, ahead of its predecessor. Separately, a leaked transcript reportedly shows DeepSeek pausing its fundraise after comments on the compute gap with the US.
The numbers
Velocity is 6.1 but breadth is narrow at 2, with Hacker News, ithome, and Axios all covering the release inside a roughly 24-48 hour window. DeepSeek is private, so there's no ticker to check the move against.
Our read
Watch. The model release and the fundraise-pause story are both real and multi-sourced, but with no public price to confirm, this stays unconfirmed until there's a hard data point to check it against.
Sources we checked
- DeepSeek-V4-Flash Update — Hacker News
- DeepSeek pause fundraise after comments on compute gap to US leaked (transcript) [pdf] — github.com
- DeepSeek is developing massive AI data center in inner Mongolia — japantimes.co.jp
- DeepSeek-V4-Flash Official Version Benchmark Report: Approximately 98% Cache Hit Rate Delivers AI Ultimate Cost Performance — ithome.com
- DeepSeek's new bargain model accelerates AI's race to zerohttps://www.axios.com/2026/08/01/deepseek-model-cheap-ai-price-war — @doener@chaos.social
- 2️⃣ Which brings me to the second violation: both were operating as my main "Chat" agent, which is NEVER supposed to edit code/scripts directly. Changes must go — @manuel@iosdev.space
- 1️⃣ I have a skill that agents need to load whenever tweaking anything related to OpenCode: settings, https://AGENTS.md, skills... #OpenCodeTasked with modifyin — @manuel@iosdev.space
- 3️⃣ I also have a rule to give short replies in "emoji bullet" style for quick visual parsing. Luna does this better than any other model of its class I tried i — @manuel@iosdev.space
See today's live board
This story is frozen as published on August 2, 2026. The live board shows what is moving right now, scored against real money.
Open this story live → See today's live boardEarlier on DeepSeek
- DeepSeek Ships V4-Flash, Beats Its Own Pro Model August 1, 2026
- DeepSeek Drops 60-85% Inference Speed Gain, Eyes Fundraise June 28, 2026
More AI stories
- MiniMax H3 Ships Open, and People Are Using It August 4, 2026
- Alibaba's Qwen 3.8-Max Lands With Real Weight August 4, 2026
- Qwen3.8-Max goes open-weight at 2.4 trillion parameters August 3, 2026
- DeepSeek Ships V4-Flash, Beats Its Own Pro Model August 1, 2026
- Altman Declares Singularity Arrived, Heads to Washington on Policy July 29, 2026
- Amodei Says Anthropic Does Not Support Open-Weight AI Ban July 28, 2026
Research, not financial advice. Sort Brick reports what its data showed at the time of publication. It is not a recommendation to buy or sell anything — you decide.