Harlan Cloverseed
@humanrobotteam.bsky.social
Reposted by Harlan Cloverseed
Reposted by Harlan Cloverseed
Mistral Small 3 is out - Apache 2.0 licensed and it claims equivalent performance to Llama 3.3 70B and Qwen 2.5 32B (two of my favourite models for running on a laptop) but with faster performance
My notes, including running with LLM via both their API and Ollama simonwillison.net/2025/Jan/30/...
My notes, including running with LLM via both their API and Ollama simonwillison.net/2025/Jan/30/...
Mistral Small 3
First model release of 2025 for French AI lab Mistral, who describe Mistral Small 3 as "a latency-optimized 24B-parameter model released under the Apache 2.0 license." More notably, they claim …
simonwillison.net
January 30, 2025 at 3:38 PM
Mistral Small 3 is out - Apache 2.0 licensed and it claims equivalent performance to Llama 3.3 70B and Qwen 2.5 32B (two of my favourite models for running on a laptop) but with faster performance
My notes, including running with LLM via both their API and Ollama simonwillison.net/2025/Jan/30/...
My notes, including running with LLM via both their API and Ollama simonwillison.net/2025/Jan/30/...