Agentic market $10.8B and climbing  ·  editor@gaasnews.com
Sections
HomeWhat is GaaS?PlatformsPricingGlossaryOpinionAboutContact
HomePricing & ModelsGemini 3.5 Slip
Pricing & Models

The Gemini 3.5 Pro slip is redrawing the price war Google was supposed to win

Bloomberg reported the flagship is months behind on coding performance, Alphabet shed roughly $200 billion in value, and China's trillion-parameter launches poured into the gap this weekend.

AJ
Andrew Jamerson
Founding Editor
Jul 19, 2026 · 4 min read
Illustration: a gap in the lineup. // GaaS News
TL;DR
  • Gemini 3.5 Pro is months behind schedule because coding and complex reasoning fell short of internal targets, Bloomberg reported Thursday, and Alphabet shares fell about 4 percent, erasing roughly $200 billion in value.
  • Google previewed the model at I/O in May with rollout expected around June; there is still no model card, pricing page, or benchmark disclosure.
  • The weekend's Chinese launches, Kimi K3 and Qwen3.8 Max, poured directly into the opening, resetting the price-performance ladder while Google's flagship slot sits empty.

The most consequential pricing event of the weekend was a product that did not ship. Gemini 3.5 Pro, previewed at Google I/O in May with broad rollout expected around June, is months behind schedule because coding performance and complex long-horizon reasoning fell short of internal targets, Bloomberg reported Thursday, with CNBC confirming. Alphabet fell about 4 percent on the news, a haircut that secondary coverage put at roughly $200 billion in market value. As of this weekend Google has published no model card, no pricing page, and no benchmarks.

Every price war has a pace-setter, and Google was supposed to be it. Gemini's flagship tier has historically anchored the aggressive end of the agent-model price ladder we have been tracking, forcing OpenAI and Anthropic to respond on cost per token. A missing flagship does not pause that war. It moves the front line.

The gap got filled in 72 hours

Consider what poured in while the slot sat empty. Moonshot's Kimi K3, at 2.8 trillion parameters with open weights due July 27, listed $3 per million input tokens and $15 per million output. Alibaba previewed the 2.4 trillion parameter Qwen3.8 Max on Sunday with an open-weight release promised. OpenAI's GPT-5.6 Sol pitch is token efficiency, with CEO Sam Altman claiming 54 percent fewer tokens on agentic coding tasks, per CNBC. The competitive set Gemini 3.5 Pro was specced against in May no longer exists.

What the desk is watching

Delays happen, and Google's distribution means a strong late model recovers fast. The pricing question is narrower: whether Google re-enters at the premium tier it planned, or is forced to launch into a ladder that open-weight flagships have re-priced beneath it, the dynamic that has already made cheap tokens the industry's favorite misleading metric. Months, in this market, is a geological age.

AJ

Andrew Jamerson

Founding Editor, GaaS News

Andrew Jamerson is the founding editor of GaaS News, covering the economics of the agent era. He started the publication to cover Agentic AI as a Service as a dedicated beat and edits every article on the site.

Be on the list when the beat breaks

One email when a platform ships, a round closes, or the ground shifts under the software stack.