Comment by kristianp
7 days ago
Interesting that they target around 120 billion parameters. Just enough to fit onto a single H100 with 4 bit quant. Or 128GB APU like apple silicon, AMD AI cpus or the GB spark.
Copying GPT-OSS-120b?
Available to try at https://build.nvidia.com/mistralai/mistral-small-4-119b-2603
Hopefully better than gpt-oss-120b because this sucks big time. Completely unusable. gpt-5.3 and 4 are very fine though.
Testing it tomorrow