- Bay Area Times
- Posts
- Alibaba’s Qwen, Meta’s Llama 3 lead Hugging Face’s new open weights LLM leaderboard
Alibaba’s Qwen, Meta’s Llama 3 lead Hugging Face’s new open weights LLM leaderboard
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/e17c23f3-58a2-4d15-b965-192be2687e69/Bay_Area_Times_email.png)
Brought to you by:
Top stories today:
- Alibaba’s Qwen leads Hugging Face’s open weights LLM leaderboard
- AirPods with camera modules may launch in 2026
- 51M people watched CNN debate live, skewing towards 55+
0. Data and calendar
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/90b6435a-fb78-4de7-afc9-bba435bf014a/Screenshot_2024-07-01_at_7.12.02_AM.png?t=1719828763)
All values as of 6 AM ET / 3 AM PT, other than S&P500 and NASDAQ close (4 PM ET / 1 PM PT).
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/09b34f08-4d63-4eea-9aba-f1b975f95b4b/Screenshot_2024-07-01_at_7.04.08_AM.png?t=1719828292)
All times are ET.
1. Alibaba’s Qwen, Meta’s Llama 3 lead Hugging Face’s new open weights LLM leaderboard
Benchmarks used: MMLU-Pro, GPQA, MuSR, MATH, IFEval, and BBH, as the standard benchmarks like MMLU “became too easy for models.”
June 3: when MMLU-Pro was released.
Google’s Gemma 2 is the leading open model on LMSYS’s leaderboard, where people vote (blindly) on which AI chatbot is best
2. AirPods with camera modules may launch in 2026
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/f8a3e05b-39ee-45bc-b1b7-895210fcbd8f/airpods_with_camera.png?t=1719775121)
Goal is to improve surround sound capabilities, not to take pictures.
IR camera similar to the iPhone Face ID receiver.
Apple is already working on iOS 19, macOS 16, watch OS 12, and visionOS 3.
Apple Intelligence is coming to Vision Pro, but not the HomePod.
The cheaper Vision may have a larger, lower-res screen (2 to 2.1 inch, with a pixel density of 1700 pixels per inch (PPI), vs. 1.42 inch and 3,386 PPI.
Be part of the open source rewards revolution. Join Homebrew creator Max Howell on his mission to reward open source developers for their valuable contributions.
Proof-of- Contribution (“PoC”): OSS projects are rewarded based on their impact via this novel consensus mechanism. Valuable codebases garner higher rewards.
Not a developer? Receive rewards by supporting your favourite open source projects and securing the OSS ecosystem.
Sign up for tea’s incentivized testnet today for the chance to earn rewards.
*Sponsored. Click the links to subscribe to tea's newsletter and learn more about how to get rewards for your open-source contributions.