AIToday
Large Language ModelsAI Business & IndustryTomasz Tunguz (Theory Ventures)Published: Sep 24, 2026, 06:00 JST

Anthropic cuts Opus price as OpenAI matches in 90 minutes

Anthropic cuts Opus price as OpenAI matches in 90 minutes

3 Key Points

  1. What happened

    Yesterday Anthropic released a new model and cut the Opus line to $5 & $25 per million tokens, the first move in five releases; OpenAI matched ninety minutes later.

  2. Why it matters

    Competition in the market's middle is about price, not top capability, since the priciest models draw only a thin slice of gateway spending.

  3. What to watch

    Whether the shift toward commodity tokens persists will determine the AI market's economics. Watch Fable 5.1's 3.7% gateway spending share in its first twelve days.

WHO IT HITSEnterprise AI budget owners and procurement teams get cheaper options in the middle tier, while providers of premium frontier models face pressure as corporate accounts move token consumption away from the most expensive tiers.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Context & Analysis

The price cuts cap a pattern that has played out since June. Anthropic set the frontier price at $10 & $50 per million tokens with Fable 5, OpenAI answered in July with GPT-5.6 Sol at $5 & $30, and the Opus line had stayed at $5 & $25 across Opus 4.5, 4, 4.8 and 5. Yesterday's move was the first change to that line.

The pressure is not only from closed-source rivals. Open models run a majority of token volume on the gateways that publish data, at an 86% discount to closed models' blended price. Large customers are also fine tuning open-weight bases: Cursor's Composer 2 fine tuned Kimi K2.5 to cut overall cost 86% against its previous in-house model, and Harvey cut cost per cell 55% against Sonnet 5 while scoring higher than Fable 5.

The right tail looks thinner than expected. Fable 5.1 took just 3.7% of gateway spending in its first twelve days, and among large corporate accounts frontier models fell from 53% of token consumption in early August to 45% by September. The outcome likely hinges on whether those tokens shift to commodity, since that would decide the economics of the AI market.

FAQ
How much did Anthropic cut the Opus line price to?
Anthropic cut the Opus line to $5 & $25 per million tokens. This was the first price move for the Opus line across five releases.
What share of gateway spending did the most capable model get?
Anthropic's Fable 5.1 commanded only 3.7% of gateway spending in its first twelve days. Its predecessor peaked at 13.2% in July, then fell to 4.9% a month later.
How fast did OpenAI respond to Anthropic's move?
OpenAI released a model and cut prices ninety minutes after Anthropic did. At the low end, OpenAI cut Luna by 80% in July and another 50% yesterday.
Tomasz Tunguz (Theory Ventures)Read Original Article

Get the latest Large Language Models news every morning

For example, today's edition would include:

  • Schmidhuber joins Sakana AI to lead RSI LabITmedia AI+ · 2h ago
  • TypeSafe AI launches Jev, claiming up to 200x faster inference without hallucinationsITmedia AI+ · 2h ago
  • Anthropic touts Claude's ART enzyme find; CRISPR researcher says routine genome miningTHE DECODER · 2h ago

AI-summarized, only the topics you pick: one digest a day via Email, LINE, or Slack.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleGoogle drops Gemini 3.8 TTS with 2,000 voices