pookiebear on Nostr: The Qwen3 series are the best open source pure text-to-text models. Qwen3.5 and ...
The Qwen3 series are the best open source pure text-to-text models.
Qwen3.5 and Qwen3.6 are tuned for multimodality and agentic tool usage.
Gemma models are about edge computing.
The narrative is changing towards useless bullshit.
I don't see the same knowledge and thinking density in other models. It's the most used models for ML research too I heared.
When governments realize how much valuable those models are, they'll try slowly closing the open source gateways, so I strongly advise backing up those models!
Published at
2026-05-09 20:49:46 CESTEvent JSON
{
"id": "5c3fa9fe69de3dfc4f6744d4d1af2b354af8adad27516318f92e01664fb09b48",
"pubkey": "2d7c783f8a393e19da8dc7dd11a2aee02f0f32bb0a2c9b1c81986534a78ae6cb",
"created_at": 1778352586,
"kind": 1,
"tags": [],
"content": "The Qwen3 series are the best open source pure text-to-text models. \n\nQwen3.5 and Qwen3.6 are tuned for multimodality and agentic tool usage.\nGemma models are about edge computing.\nThe narrative is changing towards useless bullshit. \n\nI don't see the same knowledge and thinking density in other models. It's the most used models for ML research too I heared. \n\nWhen governments realize how much valuable those models are, they'll try slowly closing the open source gateways, so I strongly advise backing up those models!\n",
"sig": "ba62fc2f95ddfd92ee81d1faac18ffe1a409da1022f8fb15911a574aa0cfd5d3d5652f980e4d4ec40179e4401c68caa553a76ce3b0d42ed9fb5a324f7d66f0d2"
}
