Find published research by company, product, platform or technology.
Showing 1–20 of 70 dossiers
CS-4 combines three WSE-3 Turbo wafers with Cerebras’ Nexus rack design. The practical shift is architectural: compute, power and I/O become modular, while Cerebras now says the same platform is intended to support CS-5 in 2027 and a 3D-memory CS-6 generation after that.
Groq 3 LPX is moving from architecture announcement to manufactured infrastructure. Artificial Analysis measured about 3,400 output tokens/s at both 10K and 100K context on an NVIDIA-hosted private endpoint, but the single-concurrency benchmark does not yet establish public-cloud price, multi-tenant throughput or end-to-end agent speed.
Jalapeño is now working first-party silicon rather than a roadmap item. OpenAI reports materially better latency and throughput per kilowatt than compared Blackwell systems across GPT-OSS, DeepSeek and Kimi workloads, while SemiAnalysis says it inspected the chip and benchmarked it with its open InferenceX suite.
Meta has made the privacy-versus-price trade explicit in its Model API: developers can choose standard pricing or a contributor model ID with steeply discounted inference in exchange for training-data permission. The choice matters for proprietary code, customer data and AI SaaS workloads.
SnapStart previously covered only selected managed runtimes; extending it to container images changes the latency-versus-packaging trade-off for teams shipping large dependencies or standard container bases, with regional exclusions and runtime-specific guidance still applying.
Self-Hosted Machines changes the architecture of Cursor’s Cloud Agents more than another model option would. Teams can keep code, build outputs, secrets and terminal/browser actions on infrastructure they control, but the planning/inference loop remains a Cursor service and enterprise teams become responsible for worker images, scaling, secrets and production validation.
GitHub Spark stops being available to existing users on August 31, 2026. Deployed apps are meant to keep running, but owners should export code to a repository now; Spark apps using `llm()` need a separate inference provider because the underlying GitHub Models service retired July 30.
Gemini 3.8 Flash keeps 3.7 Flash’s promotional per-token rate and Flash-tier latency, but early independent analysis suggests harder reasoning can increase tokens consumed per task. A separate 3.8 Flash Cyber model is available only through Google’s Fairwind defensive-security program.
Gemini 3.5 Transcribe turns Google’s audio understanding into a purpose-built developer surface: low-latency live transcription costs roughly $0.009/minute at Google’s published assumptions, while file transcription is roughly $0.005/minute and supports richer metadata.
Grafana’s GA agent-observability stack can track latency, tokens, cost and conversations, score live traffic with deterministic or LLM-based evaluators, route failures into test collections, compare experiments and use those results as pull-request gates. Evaluator quality and instrumentation coverage remain the main limits.
Project Zenith is not a new model or another Copilot feature. It standardizes a developer-focused Windows experience and hardware floor for local AI work, with preconfigured tooling and OS settings intended to reduce setup friction and dependence on metered cloud inference.
Hugging Face has released 207 Apache-2.0 WebGPU kernels, a JavaScript loader and Fleet, a browser benchmarking service. The package makes kernel contracts and correctness evidence inspectable, but performance remains device- and workload-dependent.
AgentControl now spans more production stacks: applications can resolve different prompts and models by context, track token/cost behavior, require approvals, use Bedrock without proxying inference through LaunchDarkly, and inspect multi-step agent runs as one conversation.
The price changes are not uniform: H100/H200 rise about 14%, B200 30%, B300 25% and GB300 about 11%. Builders using dedicated inference or training should re-run workload economics before assuming newer accelerators remain the cheapest route per completed task.
Sentence Transformers 6 now has both unified multi-vector inference and a documented end-to-end training workflow. A new project-authored benchmark shows fast domain adaptation on a single GPU, but the result is workload-specific and index costs remain high.
Google’s new agent FinOps model combines hard monthly spend caps that pause agent API calls, Flexible Savings Plans with one- or three-year commitments, pay-as-you-go Gemini Enterprise usage and planned deferred execution at up to half normal inference cost. The controls are useful, but commitment economics and task eligibility need to be modeled carefully.
DeepSeek V4 Pro combines a production model release with peak/off-peak API pricing: cached input, uncached input and output all cost 50% less outside two daily peak windows. Builders running deferrable workloads can now treat scheduling as part of model-routing economics.
Muse Voice Transcribe gives voice-app builders one streaming model for transcription, speaker separation and turn detection instead of stitching those stages together. Its low published price is notable, but Meta’s benchmark claims still need workload-specific validation.
DigitalOcean’s MySQL 8.0 support window now has a hard operational endpoint. Existing managed clusters need application and schema compatibility testing before October 30 because the provider will move them to 8.4 during maintenance rather than leave 8.0 running indefinitely.
Aurora Serverless can now add roughly 12 ACUs in the first second of a scale-up event on platform versions 3 and 4. The change is automatic and is most useful for bursty SaaS, API, batch and agent workloads, but it does not remove the separate resume delay when a database has scaled all the way to zero.