Model: "qwen2.5-32b-q3_k_m"

Nov 19, 2024
Pixtral Large (124B) beats Llama 3.2 90B with updated Mistral Large 24.11
Show details
pixtral-large mistral-large-24.11 llama-3-2 qwen2.5-7b-instruct-abliterated-v2-gguf qwen2.5-32b-q3_k_m vllm llama-cpp exllamav2 tabbyapi mistral-ai sambanova nvidia multimodality vision model-updates chatbots inference gpu-optimization quantization performance concurrency kv-cache arthur-mensch
Mistral has updated its Pixtral Large vision encoder to 1B parameters and released an update to the 123B parameter Mistral Large 24.11 model, though the update lacks major new features. Pixtral Large outperforms Llama 3.2 90B on multimodal benchmarks despite having a smaller vision adapter. Mistral's Le Chat chatbot received comprehensive feature updates, reflecting a company focus on product and research balance as noted by Arthur Mensch. SambaNova sponsors inference with their RDUs offering faster AI model processing than GPUs. On Reddit, vLLM shows strong concurrency performance on an RTX 3090 GPU, with quantization challenges noted in FP8 kv-cache but better results using llama.cpp with Q8 kv-cache. Users discuss performance trade-offs between vLLM, exllamav2, and TabbyAPI for different model sizes and batching strategies.