Costco Wholesale has for
Members: 13.6" Apple MacBook Air Laptop w/ M4 Chip (2025, Various Colors) on sale for
$999.97.
Shipping is free.
Thanks to Community Member
mercedes03 for sharing this deal.
- Note: Availability may vary by location. Our links may go to a product variant that is out of stock or not included in this deal. Please use the sorting options on the product page to select the items on sale.
Available:
- Midnight (MC6C4LL/A) $999.97
- Silver (MC654LL/A) $999.97
- Sky Blue (MC6V4LL/A) $999.97
Specs:
- 13.6" 2560 x 1664 Liquid Retina Display
- Apple M4 10-Core Chip
- 10-Core GPU | 16-Core Neural Engine
- 24GB RAM
- 512GB SSD
- Wi-Fi 6E (802.11ax) | Bluetooth 5.3
- FaceTime Full HD 1080p Camera
- Backlit Magic Keyboard
- Force Touch Trackpad | Touch ID Sensor
- 35W Dual USB-C Compact Power Adapter
- MacOS
- 2.7 lb
- Ports:
- 2x Thunderbolt 4
- 1x MagSafe 3
- 1x 3.5mm Headphone
Top Comments
Community Reviews
This 13.6 inch MacBook Air stands out as a strong everyday laptop with 24GB of unified memory and a 512GB SSD, which makes it especially appealing for students, heavy multitasking, and future proofing. The extra RAM is the main draw here, since it cannot be upgraded later and helps with browser tabs, Xcode or Android Studio, virtual machines, and some light local AI work. Costco also adds protection perks and an extra warranty. The main tradeoff is that the newer M5 alternatives offer faster chip features for only a bit more, while this listing has already sold out in some colors.
People running local AI models or VMs on a portable laptop, because 24GB can fit larger models and memory is the main bottleneck.
Students and office workers who keep lots of tabs open, because 24GB RAM helps avoid swapping as apps and browser use grow over time.
Average everyday laptop buyers who want a long-lasting machine, because this configuration is described as enough for regular tasks for years.
Students and hobby coders using Xcode or Android Studio with an emulator, because 24GB can keep the IDE and emulator running without heavy swapping.
People who want the newest chip for AI acceleration, because the M5 offers higher memory bandwidth and faster local model performance.
People expecting a laptop for sustained local LLM work, because the Air can run too hot and throttle under long loads.
Light users who only browse, email, and do Zoom calls, because 24GB is overkill for those workloads.
Buyers who need a machine that can be upgraded later, because the RAM cannot be added after purchase.
Common Questions (5)
AI-generated summary based on comments from the community. Updated ago
Join The Conversation
Share your experience with the Slickdeals community
Share information with the community. Please follow our Community Guidelines and be kind!
79 Comments
Sign up for a Slickdeals account to remove this ad.
i initially thought it was the M5
all that said: get the most memory you can now. and check benchmarks by googling "omlx benchmark" in advance. tough choice, i was going to say "get 24GB with M4" but the harsh truth is that M5 might have huge uplift for models.
keep in mind, ive heard they burn your battery. never tried myself.
Sign up for a Slickdeals account to remove this ad.
Last summer I built an AI rig just to dabble...96gb was 162.99 (ah, the good old days) and I figured that it would hurt to overbuild considering that the parts were cheap. I also got a used 3090fe for around $580 a few months prior. Under a $1200 out the door for my first rig since 2015.
Starting this year, I had to write a new sat curriculum from total scratch and got serious about putting the machine's muscle to use. I've been using qwen3 14b (have felt no need to go to 3.5 yet) at Q4, and it's been great at showing me alternate explanations for many questions that students ask. I originally tried the 32b, but it only allowed for 8k context, and if I went to Q3, it would start literally rounding numbers the wrong direction...so I lowered the parameters, and anted up the window to 32k, and been flying ever since. Haven't dabbled with any local AI agents yet (which you shouldn't do with your primary from what I've read), but these m4 machines seem to be flying off of the shelf for things like OpenClaw.
Give yourself some room to grow, and you'll definitely find a use case if you have interest in trying something new in the world of AI.
Just bear in mind that some models have more conservative sweet spots than others, so while several might 'fit' in a machine's memory, your usage case might demand something different than other people. Some people want speedy generation (for real-time answers to questions), others want a large context window (so you can paste in a whole book, code, or have a long convo), or have a large parameter model (for more nuanced language).
Pick M5 + 16 GB if your work is web browsing, email, Zoom, office apps, light photo editing, music software with moderate track counts, or you want better longevity from a newer chip.
Pick M4 + 24 GB only if you regularly run memory‑heavy workflows (50+ browser tabs, big image libraries, multiple VMs, very large spreadsheets) and you know you exceed 16 GB often
Sign up for a Slickdeals account to remove this ad.
one small isse that one port doest have power flow throught it or something, never really bothered since i dont use it.
anythoughts?
Join The Conversation
Share your experience with the Slickdeals community
Share information with the community. Please follow our Community Guidelines and be kind!