DeepSeek's V4 Test Release Ignites LLM Price War

DeepSeek released the test version of its V4 model, sparking widespread discussion in the AI industry due to its outstanding performance and extremely low cost. Multiple industry observers believe this will reshape commercial expectations for the large model market, marks the beginning of a full-scale price war, and could even "make history."

Confirmed

Model Release: According to Nikkei Asia, DeepSeek has released a test version of its V4 model.
Price War Initiated: Gary Marcus pointed out that this event marks the full-scale AI price war he had repeatedly predicted since August 2023. Previously, investor Michael Burry had also issued similar warnings about an industry price war.

Unconfirmed

Competitor Details: Discussions mentioned advanced OpenAI models (like GPT-5.6 Luna), but specific details and positioning of these competing products only exist in user comparisons.

Why it matters

Compressing Profit Margins: teortaxesTex relayed jmbollenbacher's view that DeepSeek's low cost makes OpenAI's pricing appear excessively profitable, indicating that profit margins for LLM APIs are about to be significantly compressed.
Accelerating Market Elimination: rickasaurus noted that under the dual pressure of performance and cost-effectiveness, most models globally have entered DeepSeek's "kill zone," and models with poor capabilities and high prices are facing elimination.
Triggering Jevons Paradox: teortaxesTex believes DeepSeek possesses an extraordinary signal-to-noise ratio (SNR) and perfectly combines open-source nature with powerful capabilities. This destructiveness could trigger the "Jevons Paradox," completely changing the ARR landscape of domestic tech giants.

2026-08-01 ~ 2026-08-01 · 6 related posts

Primary sources

DeepSeek Releases V4 Beta as AI Price War Intensifies — GaryMarcus · 2026-08-01 03:57 (UTC+08:00)
DeepSeek V4 Flash's low price sparks debate: OpenAI margins called excessive — teortaxesTex · 2026-08-01 02:33 (UTC+08:00)
DeepSeek's Upcoming Model Sparks 'Jevons Paradox' Predictions for Domestic AI Market — teortaxesTex · 2026-08-01 04:32 (UTC+08:00)

[source] DeepSeek V4 Flash's low price sparks debate: OpenAI margins called excessive — teortaxesTex · 2026-08-01
[source] DeepSeek Releases V4 Beta as AI Price War Intensifies — GaryMarcus · 2026-08-01
[source] DeepSeek's Upcoming Model Sparks 'Jevons Paradox' Predictions for Domestic AI Market — teortaxesTex · 2026-08-01
KOL Praises DeepSeek: High SNR and Open Prowess to Trigger Jevons Paradox — teortaxesTex · 2026-08-01
DeepSeek's Price Competitiveness Threatens OpenAI's High Margins — teortaxesTex · 2026-08-01
DeepSeek's Price-Performance Ratio Forces Inferior, Expensive Models Out of the Market — rickasaurus · 2026-08-01