- The Upload
- Posts
- OpenAI Just Dropped the Price
OpenAI just made its voice AI way cheaper—and way more fun. But that’s not all: they also dropped a high-priced model, Claude finally got web browsing, and Google’s Canvas is shaping up to be a new creative playground.
🔍 Claude finally gets real-time web search
🎙️ OpenAI’s voice AI gets a vibe upgrade (and a major price drop)
💰 OpenAI’s new o1-Pro model is 10x the price
✏️ Gemini’s Canvas gives you a smarter space to create, code, and collaborate
Let’s get into it. 🚀
Claude can now search the web.
Each response includes inline citations, so you can also verify the sources.
— Anthropic (@AnthropicAI)
4:51 PM • Mar 20, 2025
🔍 Claude finally gets real-time web search
Anthropic just flipped the switch on Claude’s web search tool, finally giving its flagship model access to the real-time internet.c
Key Details:
Claude 3.7 Sonnet now searches the web on demand and includes source citations.
The feature is available to paid U.S. users, with global rollout coming soon.
Toggle it on in your settings under "Web Search."
Why it matters:
Claude has long been a top-tier model—but without browsing, it was running blind. This update brings it closer to parity with ChatGPT’s search and Perplexity’s research mode.
What it means for you:
If you rely on Claude for research or writing, you can now get live, cited info without leaving the chat. Welcome to the future.
Image source: OpenAI
🎙️ OpenAI’s voice AI gets a vibe upgrade (and a major price drop)
OpenAI launched new speech-to-text and text-to-speech models with customizable speaking styles—and they're coming for ElevenLabs’ lunch.
Key Details:
Style prompts like “bedtime story voice” or “pirate” control tone, pace, and emotion.
The new model,
gpt-4o-mini-tts, is 85% cheaper than ElevenLabs at ~$0.015/min.OpenAI.fm is the new playground for testing styles and vibes.
Why it matters:
OpenAI’s voice tech still isn’t as emotionally rich as rivals, but this update closes the gap—and the pricing could seriously shake the market.
What it means for you:
Whether you're a marketer, podcaster, or builder, voice content just got faster, cheaper, and more customizable.
Image source: OpenAI
💰 OpenAI’s new o1-Pro model is 10x the price—what’s it for?
OpenAI dropped its o1-Pro model via API at a staggering $150–$600 per million tokens, compared to just ~$15–$30 for standard models.
Key Details:
o1-Pro is aimed at enterprise users with demanding performance needs.
Likely optimized for reliability, latency, or model alignment, but OpenAI hasn’t fully explained the premium.
The pricing far exceeds Claude Opus and GPT-4 Turbo.
Why it matters:
This might signal a move toward "luxury-tier" AI models for highly specific, regulated, or sensitive use cases.
What it means for you:
Unless you're running mission-critical operations, this probably isn’t for you—but it's a big signal that AI pricing tiers are about to diversify.
✏️ Gemini’s Canvas gives you a smarter space to create, code, and collaborate
Google’s Gemini now includes Canvas, a visual workspace for brainstorming, coding, and creative testing.
Key Details:
It’s like a mix of ChatGPT’s code interpreter, Claude’s artifacts, and Figma’s boards.
You can generate visuals, edit code, and manipulate layouts in real-time.
Great for side projects, marketing visuals, or dev prototyping.
Why it matters:
Everyone’s copying everyone—but Canvas makes Gemini feel more tactile and interactive, giving users more control over ideas and outputs.
What it means for you:
If you’re a content creator or strategist, Canvas gives you a real space to think and create—with AI actively co-building alongside you.
🎯 Key Takeaways
Claude finally has real-time web search, putting it on par with ChatGPT and Perplexity.
OpenAI’s new voice AI is cheaper and more expressive, signaling major shifts in audio content creation.
OpenAI’s o1-Pro shows premium models are coming—but at steep prices.
Gemini’s Canvas adds creative flow to AI, helping users co-build smarter and faster.