Tag
A community-compiled ranking of 76 GPUs benchmarked with llama.cpp's Vulkan backend for local LLM inference, sorted by price-to-performance on the used market. Older wide-bus and HBM2 cards like the P102-100, MI50, and Radeon VII dominate value, while modern mid-range cards are bottlenecked by narrow memory buses.
The article discusses AI models ranked on Pareto frontiers across benchmark categories, highlighting how Claude Sonnet 5.5 has improved Anthropic's position and pushed some OpenAI models off the frontiers in terms of quality, price, and speed.
The tweet speculates on the potential release of GPT-6 Sol and Luna, which could significantly improve price/performance and lower costs for advanced AI intelligence.
Ox Alpha, based on GLM-5.3 Flash AA, is announced at one-hundredth the price of frontier models, powered by Chinese chips, and has achieved nearly 20% weekly token share on OpenRouter.
Modular demonstrates strong price-performance on Zai_org's GLM-5.2 model in a benchmark, achieving near-top speed without high costs.
OpenAI's GPT-5.6 model family is now available in Kiro, a software development agent, offering improved price-performance for AI-native coding workflows.
Meta has globally released Muse Spark 1.2 Contributor at a discount, offering better price/performance than OpenAI's Luna and DeepSeek, with user data used for training but potential open weights advantages.
OpenAI introduces GPT-5.6, a new model family that delivers frontier-level agent performance at a fraction of the cost, with new API primitives for reasoning persistence, multi-agent orchestration, and programmatic tool calling.
Theo praises Luna's price-performance after an 80% cost reduction, noting it's nearly free and capable of real data processing, and plans to use it more in T3 Code for title generation.
DeepSeek-V4-Flash-High tops the Frontend Code Arena with a 1586 score, offering the best performance-per-dollar at $0.14/$0.28 per MToken.
Cognition updated FrontierCode 1.1 to include new discounts for GPT-5.6 Terra and GPT-5.6 Luna, placing the GPT-5.6 series on the Pareto curve of price/performance efficiency.
A tweet highlighting Google's Gemini 3.5 Flash as a fast, capable, and affordable AI model release, emphasizing its impressive benchmarks and price/performance ratio.