▲ 79 ▼ OpenAI announces new slop generator GPT-6 Astra: Unfortunately not any better than the last one (openai.com) submitted 3 days ago* (last edited 3 days ago) by potato_lemon@feddit.nl to c/technology@lemmy.world 37 comments fedilink hide all child comments
[–] SaharaMaleikuhm@feddit.org 27 points 3 days ago (2 children) 2 months until China drops a better one for free permalink fedilink source hideshow 4 child comments replies: [–] Mwa 14 points 3 days ago (1 child) And it runs locally permalink fedilink source parent hideshow 2 child comments replies: [–] magnue@lemmy.world 15 points 3 days ago (1 child) (on 1TB memory) permalink fedilink source parent hideshow 2 child comments replies: [–] Mwa 6 points 3 days ago* (3 children) 27B ;) or lower depends on your ram permalink fedilink source parent hideshow 6 child comments replies: [–] VibeSurgeon@piefed.social 7 points 3 days ago (2 children) I mean 27b models are practically speaking useless at anything but turning your computer into a space heater, in my experience, but yeah permalink fedilink source parent hideshow 4 child comments replies: [–] notSys@lemmy.cafe 2 points 2 days ago Qwen 3.8 27B is very good at agentic work permalink fedilink source parent [–] Mwa 2 points 3 days ago atleast they are kinda decent permalink fedilink source parent [–] PushButton@lemmy.world 3 points 3 days ago (1 child) I made the aqwentance of this one already. The new one is particularly good. permalink fedilink source parent hideshow 2 child comments replies: [–] Mwa 3 points 3 days ago i didnt test the 27B models cause of my RAM/VRAM,but they sound decent. permalink fedilink source parent [–] magnue@lemmy.world 2 points 3 days ago (2 children) Yeah but not comparable to a frontier model permalink fedilink source parent hideshow 4 child comments replies: [–] partofthevoice@lemmy.zip 3 points 3 days ago (1 child) The head of IT at my org is dying on this hill. He says specialized models with proper routing can outperform the big boys. … I really want him to be right, but I don’t believe it. I work with both. The 27B models are like working with ChatGPT on release day. permalink fedilink source parent hideshow 2 child comments replies: [–] theneverfox@pawb.social 2 points 2 days ago Well of course they outperform them if they're specialized to a purpose... Theres a like 3B model that plays Minecraft and generally kicks the ass of all the more general llms And in general, the big models are primarily better at remembering instructions and context... An 8B model can hold a basic conversation or perform simple tasks on a similar level to a frontier model Really, really depends on what they're specialized to do though permalink fedilink source parent [–] Mwa 2 points 3 days ago* better then nothing i guess,but yeah true permalink fedilink source parent [–] VibeSurgeon@piefed.social 9 points 3 days ago Chinese models are typically only cheaper than models from western labs, not higher performance. Which is better in one aspect, but not the other permalink fedilink source parent
[–] Mwa 14 points 3 days ago (1 child) And it runs locally permalink fedilink source parent hideshow 2 child comments replies: [–] magnue@lemmy.world 15 points 3 days ago (1 child) (on 1TB memory) permalink fedilink source parent hideshow 2 child comments replies: [–] Mwa 6 points 3 days ago* (3 children) 27B ;) or lower depends on your ram permalink fedilink source parent hideshow 6 child comments replies: [–] VibeSurgeon@piefed.social 7 points 3 days ago (2 children) I mean 27b models are practically speaking useless at anything but turning your computer into a space heater, in my experience, but yeah permalink fedilink source parent hideshow 4 child comments replies: [–] notSys@lemmy.cafe 2 points 2 days ago Qwen 3.8 27B is very good at agentic work permalink fedilink source parent [–] Mwa 2 points 3 days ago atleast they are kinda decent permalink fedilink source parent [–] PushButton@lemmy.world 3 points 3 days ago (1 child) I made the aqwentance of this one already. The new one is particularly good. permalink fedilink source parent hideshow 2 child comments replies: [–] Mwa 3 points 3 days ago i didnt test the 27B models cause of my RAM/VRAM,but they sound decent. permalink fedilink source parent [–] magnue@lemmy.world 2 points 3 days ago (2 children) Yeah but not comparable to a frontier model permalink fedilink source parent hideshow 4 child comments replies: [–] partofthevoice@lemmy.zip 3 points 3 days ago (1 child) The head of IT at my org is dying on this hill. He says specialized models with proper routing can outperform the big boys. … I really want him to be right, but I don’t believe it. I work with both. The 27B models are like working with ChatGPT on release day. permalink fedilink source parent hideshow 2 child comments replies: [–] theneverfox@pawb.social 2 points 2 days ago Well of course they outperform them if they're specialized to a purpose... Theres a like 3B model that plays Minecraft and generally kicks the ass of all the more general llms And in general, the big models are primarily better at remembering instructions and context... An 8B model can hold a basic conversation or perform simple tasks on a similar level to a frontier model Really, really depends on what they're specialized to do though permalink fedilink source parent [–] Mwa 2 points 3 days ago* better then nothing i guess,but yeah true permalink fedilink source parent
[–] magnue@lemmy.world 15 points 3 days ago (1 child) (on 1TB memory) permalink fedilink source parent hideshow 2 child comments replies: [–] Mwa 6 points 3 days ago* (3 children) 27B ;) or lower depends on your ram permalink fedilink source parent hideshow 6 child comments replies: [–] VibeSurgeon@piefed.social 7 points 3 days ago (2 children) I mean 27b models are practically speaking useless at anything but turning your computer into a space heater, in my experience, but yeah permalink fedilink source parent hideshow 4 child comments replies: [–] notSys@lemmy.cafe 2 points 2 days ago Qwen 3.8 27B is very good at agentic work permalink fedilink source parent [–] Mwa 2 points 3 days ago atleast they are kinda decent permalink fedilink source parent [–] PushButton@lemmy.world 3 points 3 days ago (1 child) I made the aqwentance of this one already. The new one is particularly good. permalink fedilink source parent hideshow 2 child comments replies: [–] Mwa 3 points 3 days ago i didnt test the 27B models cause of my RAM/VRAM,but they sound decent. permalink fedilink source parent [–] magnue@lemmy.world 2 points 3 days ago (2 children) Yeah but not comparable to a frontier model permalink fedilink source parent hideshow 4 child comments replies: [–] partofthevoice@lemmy.zip 3 points 3 days ago (1 child) The head of IT at my org is dying on this hill. He says specialized models with proper routing can outperform the big boys. … I really want him to be right, but I don’t believe it. I work with both. The 27B models are like working with ChatGPT on release day. permalink fedilink source parent hideshow 2 child comments replies: [–] theneverfox@pawb.social 2 points 2 days ago Well of course they outperform them if they're specialized to a purpose... Theres a like 3B model that plays Minecraft and generally kicks the ass of all the more general llms And in general, the big models are primarily better at remembering instructions and context... An 8B model can hold a basic conversation or perform simple tasks on a similar level to a frontier model Really, really depends on what they're specialized to do though permalink fedilink source parent [–] Mwa 2 points 3 days ago* better then nothing i guess,but yeah true permalink fedilink source parent
[–] Mwa 6 points 3 days ago* (3 children) 27B ;) or lower depends on your ram permalink fedilink source parent hideshow 6 child comments replies: [–] VibeSurgeon@piefed.social 7 points 3 days ago (2 children) I mean 27b models are practically speaking useless at anything but turning your computer into a space heater, in my experience, but yeah permalink fedilink source parent hideshow 4 child comments replies: [–] notSys@lemmy.cafe 2 points 2 days ago Qwen 3.8 27B is very good at agentic work permalink fedilink source parent [–] Mwa 2 points 3 days ago atleast they are kinda decent permalink fedilink source parent [–] PushButton@lemmy.world 3 points 3 days ago (1 child) I made the aqwentance of this one already. The new one is particularly good. permalink fedilink source parent hideshow 2 child comments replies: [–] Mwa 3 points 3 days ago i didnt test the 27B models cause of my RAM/VRAM,but they sound decent. permalink fedilink source parent [–] magnue@lemmy.world 2 points 3 days ago (2 children) Yeah but not comparable to a frontier model permalink fedilink source parent hideshow 4 child comments replies: [–] partofthevoice@lemmy.zip 3 points 3 days ago (1 child) The head of IT at my org is dying on this hill. He says specialized models with proper routing can outperform the big boys. … I really want him to be right, but I don’t believe it. I work with both. The 27B models are like working with ChatGPT on release day. permalink fedilink source parent hideshow 2 child comments replies: [–] theneverfox@pawb.social 2 points 2 days ago Well of course they outperform them if they're specialized to a purpose... Theres a like 3B model that plays Minecraft and generally kicks the ass of all the more general llms And in general, the big models are primarily better at remembering instructions and context... An 8B model can hold a basic conversation or perform simple tasks on a similar level to a frontier model Really, really depends on what they're specialized to do though permalink fedilink source parent [–] Mwa 2 points 3 days ago* better then nothing i guess,but yeah true permalink fedilink source parent
[–] VibeSurgeon@piefed.social 7 points 3 days ago (2 children) I mean 27b models are practically speaking useless at anything but turning your computer into a space heater, in my experience, but yeah permalink fedilink source parent hideshow 4 child comments replies: [–] notSys@lemmy.cafe 2 points 2 days ago Qwen 3.8 27B is very good at agentic work permalink fedilink source parent [–] Mwa 2 points 3 days ago atleast they are kinda decent permalink fedilink source parent
[–] notSys@lemmy.cafe 2 points 2 days ago Qwen 3.8 27B is very good at agentic work permalink fedilink source parent
[–] PushButton@lemmy.world 3 points 3 days ago (1 child) I made the aqwentance of this one already. The new one is particularly good. permalink fedilink source parent hideshow 2 child comments replies: [–] Mwa 3 points 3 days ago i didnt test the 27B models cause of my RAM/VRAM,but they sound decent. permalink fedilink source parent
[–] Mwa 3 points 3 days ago i didnt test the 27B models cause of my RAM/VRAM,but they sound decent. permalink fedilink source parent
[–] magnue@lemmy.world 2 points 3 days ago (2 children) Yeah but not comparable to a frontier model permalink fedilink source parent hideshow 4 child comments replies: [–] partofthevoice@lemmy.zip 3 points 3 days ago (1 child) The head of IT at my org is dying on this hill. He says specialized models with proper routing can outperform the big boys. … I really want him to be right, but I don’t believe it. I work with both. The 27B models are like working with ChatGPT on release day. permalink fedilink source parent hideshow 2 child comments replies: [–] theneverfox@pawb.social 2 points 2 days ago Well of course they outperform them if they're specialized to a purpose... Theres a like 3B model that plays Minecraft and generally kicks the ass of all the more general llms And in general, the big models are primarily better at remembering instructions and context... An 8B model can hold a basic conversation or perform simple tasks on a similar level to a frontier model Really, really depends on what they're specialized to do though permalink fedilink source parent [–] Mwa 2 points 3 days ago* better then nothing i guess,but yeah true permalink fedilink source parent
[–] partofthevoice@lemmy.zip 3 points 3 days ago (1 child) The head of IT at my org is dying on this hill. He says specialized models with proper routing can outperform the big boys. … I really want him to be right, but I don’t believe it. I work with both. The 27B models are like working with ChatGPT on release day. permalink fedilink source parent hideshow 2 child comments replies: [–] theneverfox@pawb.social 2 points 2 days ago Well of course they outperform them if they're specialized to a purpose... Theres a like 3B model that plays Minecraft and generally kicks the ass of all the more general llms And in general, the big models are primarily better at remembering instructions and context... An 8B model can hold a basic conversation or perform simple tasks on a similar level to a frontier model Really, really depends on what they're specialized to do though permalink fedilink source parent
[–] theneverfox@pawb.social 2 points 2 days ago Well of course they outperform them if they're specialized to a purpose... Theres a like 3B model that plays Minecraft and generally kicks the ass of all the more general llms And in general, the big models are primarily better at remembering instructions and context... An 8B model can hold a basic conversation or perform simple tasks on a similar level to a frontier model Really, really depends on what they're specialized to do though permalink fedilink source parent
[–] Mwa 2 points 3 days ago* better then nothing i guess,but yeah true permalink fedilink source parent
[–] VibeSurgeon@piefed.social 9 points 3 days ago Chinese models are typically only cheaper than models from western labs, not higher performance. Which is better in one aspect, but not the other permalink fedilink source parent