Mistral AI has launched Mistral Large 4 (ML-4), a 675-billion-parameter open-weight model available on Hugging Face. This model reportedly achieves sub-second latency on a single A100 GPU and is released under a permissive research license. The launch sets a new benchmark for open-source AI capabilities, demonstrating significant scale and performance without the backing of major cloud providers. AI
IMPACT Sets a new benchmark for open-source AI, potentially accelerating enterprise adoption of large-scale models.
RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
Read on Mastodon — fosstodon.org →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →