Moonshot’s Kimi K3 beats Fable 5

Welcome back to The Shift. Let’s get straight to what matters in AI today…

Today we have:

🇨🇳 Moonshot Launches Kimi K3

🧐How to Find Out Why Your YouTube Channel Isn't Growing With vidIQ

🧠 Thinking Machines Launches Inkling

🔨Tools and Shifts you cannot miss

🇨🇳 Moonshot Launches Kimi K3

Chinese AI startup Moonshot has unveiled Kimi K3, an open-weight model that reaches new performance highs for Chinese and open-source AI while competing with leading frontier models at comparable pricing.

The Decode:

Frontier-Level Performance -  Kimi K3 features a 1M-token context window and outperforms Claude Fable 5 and GPT-5.6 Sol on benchmarks covering web research, spreadsheet tasks, frontend design, and long-form coding.

Benchmark Gains - The model scored 57 on Artificial Analysis' Intelligence Index, trailing only Claude Fable 5 (60) and GPT-5.6 Sol (59) while improving by double-digit positions over K2.6.

Long-Running Demo - In one demonstration, K3 independently spent 48 hours designing and verifying a small chip capable of running a miniature version of itself, reaching 8,700 tokens per second in simulation.

Open Release Planned - Moonshot priced K3 at $3/$15 per million tokens, matching Claude 5 Sonnet, and said the model's open weights will be released by July 27.

Kimi K3 narrows the gap between open-weight and frontier AI models while reinforcing China's growing position in open-source AI development with a high-performing model available for broader developer use.

Together with MarketBeat

7 Stocks to Buy Before the Robots Take Over

The next AI trade may not be another chatbot.

It may be surgical robots, automated warehouses, smart factories, and machine vision systems already reshaping how companies operate.

MarketBeat’s new 7 Stocks to Buy Before the Robotics Revolution report reveals seven companies positioned across the automation boom, from robot builders and AI chip leaders to machine vision providers and factory automation giants.

This is where AI gets a body.

And as labor shortages, wage pressure, and supply chain stress push more companies toward automation, these stocks could move before the robotics story becomes impossible to ignore.

The report normally sells for $29.97, but it is free for a limited time.

Send My Free Report

🧐How to Find Out Why Your YouTube Channel Isn't Growing With vidIQ

No sign-up. No guesswork. Paste your channel and get a score in seconds.

Step 1: Go to the audit tool: Head to vidiq.com/audit. Enter any YouTube channel URL, ID, or name, your own or a competitor's.

Step 2: Get scored across five signals: vidIQ grades your channel on upload consistency, SEO optimization, engagement rate, views-to-subscriber ratio, and posting recency, the exact signals YouTube's recommendation algorithm rewards.

Step 3: Find your weakest signal: Your lowest score is your biggest growth lever. A weak SEO score means great content stays hidden. A low views-to-subscriber ratio means you're not breaking out of your existing audience.

Step 4: Benchmark against competitors: Audit any public channel. Enter a competitor in your niche and see exactly where they're outscoring you and why.

Step 5: Turn the score into a plan: vidIQ maps each score to specific improvement actions, keyword data to fix SEO, posting schedules to fix consistency, and engagement tactics to push the algorithm. Free to start, no credit card required.

Audit monthly. Give changes 30 to 60 days to reflect before checking again. You can audit your channel free here.

🧠 Thinking Machines Launches Inkling

Thinking Machines Lab has unveiled Inkling, its first open-weight multimodal model, focusing on customization, efficiency, and developer flexibility instead of competing solely on benchmark rankings.

The Shift:

First Open Model - Inkling is a 975B-parameter Mixture-of-Experts model with 41B active parameters, a 1M-token context window, and native support for text, images, and audio.

Built For Customization - Developers can fine-tune Inkling through Tinker or download it from Hugging Face, with the company positioning customization as its primary advantage over raw benchmark performance.

Efficiency Controls - Inkling features a controllable thinking effort setting that balances reasoning quality with cost, matching Nvidia's Nemotron 3 Ultra on one coding benchmark while using one-third the tokens.

Developer-Focused Strategy - Thinking Machines also introduced Inkling-Small and demonstrated the model using Tinker to fine-tune itself, highlighting its emphasis on practical customization for real-world workflows.

Rather than chasing benchmark leadership, Thinking Machines is positioning Inkling as an open, customizable foundation model. The launch adds a new U.S.-based option to the open-weights ecosystem, giving developers another alternative for building specialized AI applications.

Together with Hightouch

The toughest room in advertising. Ad Studio just cleared it.

Major brands just took over a Times Square billboard, and every ad was built in Hightouch Ad Studio. Sweetgreen, HelloFresh, and Tripadvisor cleared the toughest room in advertising. Create, edit, and launch on-brand ads in a few clicks.

See what you can build

🔨AI Tools for the Shift

⚙️ Script.it – Automates workflows and business processes with AI agents, helping teams build smarter automation with less manual effort.

🛡️ AISafe Labs – Uses AI security agents to assess applications for vulnerabilities and generate SOC 2 and ISO 27001 audit-ready reports in hours.

📈 Diploria AI – Tracks how AI platforms mention your brand, measures AI search visibility, and recommends ways to improve it.

🧊 MeshRefinery – Repairs damaged 3D meshes by detecting invisible holes and artifacts, then rebuilding them into clean, watertight models.

🖋️ AI Image Text Editor – Edits text inside existing images by replacing, removing, translating, or redacting it while preserving the original design and layout.

🚀Quick Shifts

🏛️ New York is using AI to review every state law and regulation, helping officials identify outdated rules in months instead of years as the state looks to modernize government policies. 

⚾ Major League Baseball has banned teams from using generative AI apps on league-issued dugout iPads during games, preventing AI-assisted recommendations for pitch calling, substitutions, and other real-time coaching decisions. 

🎥 Google Vids can now create AI videos of you using a selfie and voice sample, while new Gemini-powered editing features let users generate and modify video clips with simple natural language prompts. 

🎮 Roblox is bringing AI game creation to its mobile app, letting users turn text prompts into playable games while using engagement-based rankings to prevent low-quality AI-generated experiences from dominating the platform. 

💼 Microsoft is reportedly coaching its sales teams to position Copilot against OpenAI, Anthropic, and Google, highlighting its in-house AI models, lower costs, and end-to-end platform as it shifts away from relying on external models. 

🧩 Prompt of the Day

Prepare for Your Performance Review Like a Promotion Conversation

Most people walk into performance reviews hoping their work speaks for itself. Top performers prepare a clear narrative, back it with evidence, and steer the conversation toward future opportunities.

Turn your review into a career milestone.

Paste the prompt: Drop this into ChatGPT, then fill in your review details.

Prompt to paste

Help me prepare for my [Insert review type and timing] based on [Insert my performance this period] and my goal of [Insert what I want from this review]. Create a structured review strategy that highlights my biggest achievements, measurable business impact, growth since the last review, and lessons learned. Anticipate difficult questions or feedback, prepare confident responses, and suggest how to transition the conversation toward a promotion, salary increase, expanded responsibilities, or career development opportunities. Include a closing statement that leaves a strong, professional impression.

🤳AI Shift of the Day

That’s all for today’s edition. See you tomorrow as we track down and get you all that matters in the daily AI Shift!

If you loved this edition, let us know how much:

Forward it to your pal to give them a daily dose of the shift so they can 👇