A technical guide details serving the 753B parameter GLM-5.3 NVFP4 model across two NVIDIA DGX Stations using vLLM, achieving a peak throughput of 607.8 tokens/sec.
Available for all U.S. Gemini app users on web and mobile, the feature allows users to link financial accounts via Plaid for deeper budgeting and subscription tracking.
Gemini 4 Hydrogen leads the proprietary field with 13.5%, while GPT Astra 6 follows at 9.5% on the same challenging benchmark.
The team optimized FSDP and Expert Parallelism on 384 GH200 GPUs to reach 1,300 tokens/s per GPU, overcoming inter-node bandwidth bottlenecks.
The round, led by CapitalG, supports LangSmith, which over 7,000 organizations use to run and govern agents.
Available initially for US ChatGPT Pro subscribers via dots, the wallet requires biometric approval for every purchase while preserving existing card rewards.
The updates simplify procurement for AWS users and ensure data residency for supported GPT-6 Luna workloads in three specific regions.
Anthropic updates Claude Gateway with spend tracking and extends fast mode to global, US, and EU Vertex regions.
The release establishes strict semantic versioning rules for the C API and expands distribution channels to include Docker, Debian, and conda-forge.
The moves mark the fourth strategic M&A transaction for Cohere this year, following acquisitions of Aleph Alpha and Reliant AI.
Available for pay-as-you-go customers, the feature uses a two-stage metering pipeline and Rating Engine to enforce caps with near real-time visibility under 10 minutes.
Perplexity releases open-weight Decider v1.3 and Flash v1.3, with v1.3 scoring 69.0 and Flash 63.6 on out-of-domain extraction tasks, surpassing GLiNER large v2.5 (33.6).
The clinical reasoning engine runs on NVIDIA-accelerated computing inside hospital firewalls to keep patient data onsite while identifying care gaps.
The framework supports weight and activation quantization for LLMs, text-to-image, and text-to-video models while maintaining inference speed.
Paid APIs provide turnkey enterprise-grade speech synthesis with SLAs, while open source models offer data sovereignty and full architectural control at the cost of compute and engineering overhead.
The service addresses brand consistency, fragmented tools, and creative control issues by embedding creative professionals who build workflows or deliver finished assets.
The agent, built on ElevenAgents, allows aspiring founders to converse with insights from the Inspiring Female Founders initiative while contributing new data to the research.
VIDRAFT has released a leaderboard that grades operational forecasts daily and a fusion model on Hugging Face.
A 164-line patch adds a missing Arm NEON kernel for llama.cpp's tiled K-quant matmul, achieving up to 6.5x faster prompt processing on Neoverse N1 without affecting generation speed.
Optimization reduces planning time from 0.6–0.8 s to 45 ms, enabling 20 Hz camera processing on a single board.
Precision Numerics releases Boole-32B-A2.5B-Preview, a 1.8-bit MoE model that is 9x smaller than BF16 and 2x faster than INT4.
OpenDocKit, a document converter supporting HWP, has been released as open source with integration for LangChain, LlamaIndex, and MCP.
The integration allows AI agents to navigate non-API systems, reducing manual technician intervention for browser-based support tasks.
A ByteDance Seed paper reveals that DeepSeek V4's chunked KV-cache compression creates periodic retrieval failures, with base model accuracy varying by 40.2 percentage points depending on token position.
AI Briefing