DeepSeek V4.1 Flash is now live on OpenRouter, with DeepSeek as the first provider. The update claims improved token efficiency and major gains in agentic benchmarks, though specific parameters and evaluation details are not disclosed.
Termsslug: token-efficiency
Token Efficiency
Token 效率
A metric measuring a model's ability to complete tasks with minimal token consumption.
Curated Intelligence
OpenRouter (@OpenRouter)T1·Models
42 pts
OpenAI 官方博客T1·Models
68 pts
OpenAI has officially released GPT-5 for enterprise use, emphasizing a 'single powerful model' experience that requires no manual selection or setup. The version sets new state-of-the-art benchmarks in math (99.6% on AIME 2025), coding (74.9% on SWEBench), and multimodal understanding, while reducing factual errors by 45% compared to GPT-4o and improving output token efficiency by 50-80% over o3.
OpenAI 官方博客T1·Models
62 pts
OpenAI introduces GPT-5.1-Codex-Max, a faster and more intelligent agentic coding model designed for Codex. It is optimized for long-running, project-scale work with enhanced reasoning capabilities and improved token efficiency.