←back to thread

Un Ministral, Des Ministraux

(mistral.ai)
216 points veggieroll | 3 comments | 16 Oct 24 14:31 UTC | HN request time: 0.001s | source
Show context
lairv ◴[16 Oct 24 14:57 UTC] No.41859815[source]▶
Hard to see how can Mistral compete with Meta, they have order of magnitude less compute, their models are only slightly better (at least on the benchmarks) with less permissive licenses?
replies(8): >>41860302 #>>41860361 #>>41860373 #>>41860530 #>>41861065 #>>41861206 #>>41861265 #>>41865550 #
simonw ◴[16 Oct 24 15:34 UTC] No.41860302[source]▶
Yeah, the license thing is definitely a problem. It's hard to get excited about an academic research license for a 3B or 8B model when the Llama 3.1 and 3.2 models are SO good, and are licensed for commercial usage.
replies(2): >>41860447 #>>41860586 #
1. harisec ◴[16 Oct 24 15:46 UTC] No.41860447[source]▶
Qwen 2.5 models are better than Llama and Mistral.
replies(1): >>41860478 #
2. speedgoose ◴[16 Oct 24 15:48 UTC] No.41860478[source]▶
I disagree. I tried the small ones but they too frequently output Chinese when the prompt is English.
replies(1): >>41860876 #
3. harisec ◴[16 Oct 24 16:16 UTC] No.41860876[source]▶
I never had this problem but i guess it depends on the prompt.