Mistral AI Launches Mixtral 8x22B: A Powerhouse for Open Weights
Mistral AI publicly releases Mixtral 8x22B, a high-performance open-weight model challenging proprietary leaders with su…
Latest articles in LLM News
Mistral AI publicly releases Mixtral 8x22B, a high-performance open-weight model challenging proprietary leaders with su…
Meta releases Llama 3 open weights, offering enterprise developers powerful, customizable AI models to compete with clos…
Anthropic releases Claude 3.5 Sonnet, setting new benchmarks in coding proficiency and reasoning capabilities for enterp…
OpenAI launches GPT-4o, a multimodal model enabling real-time voice and vision interactions with human-like latency.
Meta releases Llama 4 with significantly reduced latency and enhanced multilingual capabilities for global enterprise us…
OpenAI launches GPT-5, a major upgrade for enterprise users featuring advanced reasoning and complex problem-solving cap…
Developers report Codex's Spark model陷入 infinite loops during code analysis, wasting tokens and forcing a switch to GPT-…
Speculation mounts that Anthropic's new model leverages Alibaba's Qwen architecture through distillation techniques.
Anthropic's Claude Opus 4.8 prioritizes admitting uncertainty, but benchmarks show performance declines and potential ev…
Anthropic's new Claude Opus 4.8 mistakenly identifies as Qwen or DeepSeek in API tests, sparking 'double standard' debat…
New FK Claude proxy offers significant cost savings for developers, with rates up to 40% lower than official Anthropic p…
Xiaomi slashes MiMo-V2.5 API costs via advanced KV cache optimization, enabling near break-even pricing.