According to DeepSeek on X, the company open sourced its in-house inference framework DeepInfer, which powers DeepSeek-V4.1-Flash at scale to cut compute costs and latency.

According to AnthropicAI, Claude Haiku 5.5 runs about 75% cheaper than Haiku 4.5 while boosting speed and capability.

According to @claudeai, Haiku 5.5 is the fastest, cheapest small Claude model, cutting average run costs ~75% vs Haiku 4.5, enabling broader AI deployment.

According to @_avichawla, SIE served 4 small models on one GPU with LRU loading, cutting concurrent workload latency from 18.58s to 1.47s.

According to AnthropicAI, Claude Sonnet 5.5 runs 30% faster and costs up to 30% less than Sonnet 5, improving speed and price for most workloads.

According to @claudeai, Sonnet 5.5 runs 30% faster and costs up to 30% less than Sonnet 5, improving speed to value for AI apps.

According to CNBC... Akamai jumped after an Anthropic deal; analysts cite AI edge delivery, GPU monetization, and inference revenue upside.

According to OpenAI on X, GPT‑6 Sol and Luna cut API prices by 50% versus GPT‑5.6 promo and bring Astra strengths with faster, more efficient inference.

According to OpenAI... GPT6 Sol and Luna offer faster, cheaper models with 50% lower API prices than GPT5.6 and improved caching for scaled workloads.

According to AnthropicAI, Claude optimized 30+ bio models with GPU code, delivering 4x faster inference and open-sourced tools for broader impact.

According to emollick, a non-LLM classifier rapidly applies decision rules and runs ~10 calls per second at about $7 per hour, offering system-level gains.

According to @deepseek_ai, V4.1-Flash adds native vision, faster inference, and higher throughput, scaling to larger models for real-world deployment.

According to gdb, OpenAI unveiled GPT-6 Astra with faster multimodal inference and context upgrades, signaling new real-time agent use cases.

According to CNBC, Nvidia will buy Hugging Face for nearly $13B to integrate models, data, and inference, expanding its end to end AI platform.

According to @OpenAI, Jalapeño delivers higher throughput and lower latency per watt in one architecture without sacrificing efficiency.

According to OpenAI... GPT-5.6 Sol API and credit prices drop 20%+ for three months, signaling efficiency gains and lower unit costs for developers.

According to OpenAI... Ultrafast mode previews GPT-5.6 Sol at up to 14x speed, first via API to select customers, expanding as capacity grows.

According to @CNBC, Bank of America highlighted a lesser-known data provider as a way to play the AI boom, citing rising demand for training and inference.

According to @CNBC, Nvidia released Nemotron 3.5 Lightning as an open-source AI model to speed enterprise fine-tuning and reduce inference costs.

According to CNBC... Investors back Jensen Huang’s accelerated computing vision, spotlighting data center, inference, and software margins.