Matt Wolfe

AI News: Fable's Back But This New Model is Better?

Jul 3, 2026 29 min
aichatgptgeminiopenaiclaude
Watch on YouTube Follow Matt Wolfe on Rundown — free

Summary

AI summaries can be incomplete or wrong. Verify anything important against the original video.

Matt Wolfe discusses the redeployment of Claude Fable 5 and its new pricing, compares it to other models on Busseybench, and reviews the Gemini Spark and GPT-4 hardware.

Matt Wolfe begins by discussing the redeployment of Claude Fable 5 and its pricing, noting that it's cheaper than Opus models but not as capable. He then moves on to the Busseybench benchmark, showing how different AI models generate images of Gary Busey, highlighting the evolution of AI image generation. Finally, he explores Gemini Spark for Mac, an AI workspace that integrates various AI models and productivity tools, and touches on OpenAI's potential new hardware for Codex and Google's Gemini Spark updates. He also briefly mentions the new Gemini models and OpenAI's potential new hardware.

Verdict

Gemini Spark
AI software

Gemini Spark for Mac offers a convenient way to use various AI tools and models in one place, but some features are still experimental.

Wait

Pros

  • ✓ All-in-one AI workspace.
  • ✓ Integrates with popular tools like Google Workspace and Microsoft 365.
  • ✓ Can automate tasks across email, files, websites & more.

Cons

  • ✗ Some features are experimental.

Specs

OS macOS
Model Gemini API

Best for

  • developers
  • users who want to automate tasks

Claims & arguments

  • Claude Fable 5 Pricing

    Claude Sonnet 5 is cheaper than Opus models but less capable.

    • 23:19 Input cost is $2/million tokens, output $10/million tokens, cheaper than Opus.
    • 29:04 Performance is close to Opus 4.8 but at lower prices.
  • Busseybench Benchmark

    AI models are improving over time, shown by the benchmark results.

    • 25:12 GPT-4 is roughly on par with other models.
    • 29:18 Fable 5 is comparable to Opus 4.8.
  • Gemini Spark for Mac

    Gemini Spark is an all-in-one AI workspace that integrates various tools.

    • It can automate tasks across email, files, websites and more.
    • It integrates with Google Workspace, Microsoft 365, and other tools.
  • OpenAI's new hardware

    OpenAI is teasing new hardware related to Codex.

    • A video shows a Codex-focused device built in collaboration with Work Louder.
    • It's a square-shaped device with several buttons.
  • Google Gemini Models

    Google has released new Gemini models, Nano Banana 2 Lite and Gemini Omni Flash.

    • Nano Banana 2 Lite is faster and cheaper, with comparable performance to GPT-5.5.
    • Gemini Omni Flash is designed for high-quality video generation and conversational editing.

Key Points

  • 10:06 Claude Fable 5 is back with new pricing, costing $2 per million input tokens and $0.10 per million output tokens, making it cheaper than Opus models.
  • 22:41 Busseybench benchmark shows evolution of AI models, with GPT-4 showing significant improvements.
  • Gemini Spark for Mac offers an all-in-one AI workspace integrating various models and productivity tools.
  • OpenAI is teasing new hardware for Codex.
  • Google released Gemini Nano Banana 2 Lite and Gemini Omni Flash, focusing on speed and cost-efficiency.

Worth watching if: Anyone interested in the latest AI model updates, benchmarks, pricing, and new tools like Gemini Spark or OpenAI's potential new hardware.

Get every Matt Wolfe video extracted like this

One daily email with structured extracts of every channel you follow. Free tier covers 15 videos a month.

Sign in with Google

No credit card. Free tier forever.

Watch on YouTube