Seismora is developing an "intelligent control plane" to coordinate AI workloads across devices, edge infrastructure and cloud providers. Founder and CEO Vito Palermo told theCUBE that the goal is to spare developers from embedding routing logic in each application, with the system deciding where work runs based on capability, cost, latency and policy constraints — what the company calls "cognitive routing." He cited Stripe's reported roughly $7.5 billion deal for model-routing platform OpenRouter as evidence of growing interest in this layer.
Railway announced it is adding OpenRouter support for Priority Boarding users, giving them access to a wider range of models for agent workloads. According to the message reposted by OpenRouter, the available models include Grok 4.7 and Muse Spark 1.3, usable through Railway's Cloud Agent harness and its bots. No timing, full model list or tier details beyond Priority Boarding were given.
StepFun said Step 5 Preview has landed on OpenRouter's New & Trending leaderboard. The company thanked users for trying it and partners for bringing it into their workflows. The source gave no benchmark figures or ranking details.
OpenRouter said Inception AI's Mercury Decide is now available with a zero-data-retention (ZDR) endpoint, alongside the model's existing free endpoint. The paid ZDR endpoint is listed at $0.02 per million input tokens — 50% off the $0.04 list price — with output and cached input free and a 66K context window. OpenRouter also called Mercury Decide its fastest-growing decision model last week, a promotional claim it did not substantiate.
OpenRouter announced on X that xAI's Grok Imagine: Video 1.5 Lite is live in the Grok Imagine API and available through OpenRouter, which calls it the new workhorse for text-to-video and image-to-video. The listed pricing is $0.02 per second at 480p, $0.03 per second at 720p and $0.14 per second at 1080p. The posts do not include quality benchmarks, clip-length limits, or details on how the Lite tier differs from the full Grok Imagine video model.
On Oct 8 OpenAI added an Ultrafast service tier to the Responses API for GPT-6.1 Sol, keeping the same gpt-6.1-sol model ID and switching via service_tier; it is open to all API users but subject to separate rate limits, with global processing and US/EU data residency. OpenAI says it offers near-Astra intelligence at up to 8x the speed of Sol Standard, priced at $12 per million input tokens and $60 per million output tokens — 6x standard and above GPT-6 Astra's standard $10/$50 rates; OpenRouter said the tier is live there too. In ChatGPT, Ultrafast is limited to Codex and ChatGPT Work for the $500 Pro plan and eligible Enterprise and Edu plans. Zhidx and InfoQ reported community pushback, with users complaining that Sol Standard is already slow and the tier amounts to a price increase.
OpenRouter said on X that GPT-6 Luna is currently the fastest Decisions model, at 180ms on global requests. The post ranks Jev and Perplexity Decider behind it. No measurement methodology or latency figures for the runners-up were provided.
Third-party data and company interviews cited in the report show US enterprises moving routine AI workloads off the priciest frontier models and onto cheaper open-weight alternatives. AlphaSense found mentions of "open-weight" or "open source" models on US earnings and investor calls in August–September were six times the year-earlier level; Vercel said open models handled 56% of tokens on its AI Gateway in August versus 7% in December; and Citi data cited by Reuters put OpenRouter's open-model token share at 65% in June, up from 34% in January. Examples include AT&T running roughly 40% of AI workloads on open models with plans to reach 70% within a year, Tinder and Harvey routing routine tasks to open models, and Cursor cutting the cost of building a browser from scratch from over $10,000 with GPT-5.5 alone to $1,339 using its own Composer model plus Anthropic Opus 4.8.
Speaking at the 2026 ITValue Summit on September 17, xFusion computing vice president Wang Liangdong said AI has entered an inference-dominated phase in which global token consumption keeps surging and buyers judge infrastructure by cost per token, throughput and cluster stability rather than peak per-card performance. The article cites a range of figures: the 2026 China Development Forum said China's daily token calls topped 140 trillion in March, OpenRouter data showed China accounted for 36% of global token calls in the March 16-22 week, JPMorgan expects Chinese inference token consumption to grow roughly 370x in five years, and IDC projects annual global consumption of 150,000 Peta Tokens by 2030. It argues the 8-GPU server paradigm hits communication bottlenecks under trillion-parameter, high-concurrency inference, pointing to Huawei's Ascend 960 supernode (4,096 cards per node, 8E FP8), Lenovo's Wanquan heterogenous computing platform V5.0 and xFusion's FusionPoD for AI rack-scale designs — while warning that supernodes only suit very large workloads, interconnect standards remain unsettled, and utilization depends on sustained paying demand.
Speaking at its Stripe Tour event in Shanghai, Stripe said more than one in six signups at AI companies involve multi-account abuse, with free-trial abuse and usage-based "dine-and-dash" billing also rising. Its Radar fraud product blocked over 3.3 million high-risk free-trial attempts in 30 days across eight businesses that deployed it, the company said. Stripe also reported that the share of AI firms using hybrid pricing rose from 27% to 41% over the past year, while fixed-subscription or per-seat pricing among top AI companies fell from 40% to 24%. It said its acquisition of OpenRouter is still pending and introduced Checkout Studio, a Machine Payments Protocol, and Stripe Billing with token-based metering.
The Wall Street Journal, citing OpenRouter data, reports that business spending on OpenAI and Anthropic models was roughly even in September, compared with Anthropic holding a 75% share in January. The report says both companies need to show Wall Street they have sustainable businesses ahead of planned IPOs.
Nikkei data reported by TechNews shows 10 major Chinese AI companies — including DeepSeek, Alibaba, Tencent and Xiaomi — launched 16 new models in September, mostly lightweight and low-cost, such as DeepSeek V4.1-Flash, Zhipu's GLM-5.3-FlashX and MiniMax M3.1-Flash-Preview. BenchLM counted 20 large language models from eight major Chinese firms in July–September, up from 2 in the first quarter and 17 in the second, while the anonymous model Space Bunny Alpha — believed to be a stealth MiniMax release — surged in usage on OpenRouter. The report contrasts this with Anthropic CEO Dario Amodei's calls to slow AI R&D and crack down on illegal distillation, noting Chinese firms increasingly train and run models on domestic chips.
Perplexity has released pplx-decider-v1.1-27b, an updated open-weights multimodal decision model, now available on OpenRouter. It accepts text, JSON or image inputs and returns typed answers with probabilities. According to Perplexity Developers and OpenRouter, it costs $0.02 per million input tokens — half the price of v1 — with output free, and the company says it scores highest on the newly launched Hugging Face Decision Index 0.3 benchmark.
OpenAI’s Decisions API is now in public beta for all developers and is available through OpenRouter as GPT-6 Luna Decisions. The API lets applications choose a model, tool, or action and accept text, JSON, or images before returning typed answers with probabilities. OpenAI says the Decisions API can make decisions up to 10 times faster than GPT-6 Luna through the Responses API, though no independent measurement is provided.
OpenRouter announced on X that Cloudflare's Clef (27B) and Clef Flash (9B) open-weight decision models are now available through OpenRouter. According to the post, the models accept text, JSON or images and return typed answers with probabilities instead of generated text. Pricing is listed at $0.24 per million input tokens for Clef and $0.09 per million for Clef Flash, with output free.
Anthropic's Claude Haiku 5.5 is now officially available on Google Cloud, callable through Model Garden, according to a post from Google AI's DEV community account, and OpenRouter said the model is live on its platform too. OpenRouter described it as the first Haiku model to support reasoning effort, roughly 75% cheaper than its predecessor, measuring more than 100 tokens per second while citing a clear benchmark improvement without giving figures. Google's post calls Haiku 5.5 the fastest and most efficient member of the Claude 5.5 family, built for high-volume, latency-sensitive agentic work.
claude-mem is an open-source persistent memory plugin that captures an agent's tool-use observations during sessions, compresses them into semantic summaries with AI, and injects relevant context into later sessions; it advertises support for Claude Code, OpenClaw, Codex, Gemini, Copilot, OpenCode and more. It installs with a single npx command, stores data in a local SQLite database with Chroma-based hybrid semantic and keyword search, and lets users bring their own OpenRouter or Gemini key, fall back to an Anthropic plan, or subscribe to its hosted memory service (the project claims a 30-day free trial). The repository appeared on GitHub Trending; the usage-gain and trial claims come from the project's own README and are not independently verified.
According to The Information, Nvidia learned in July that AI model marketplace OpenRouter was being pursued for acquisition and told its leadership it was prepared to make a generous offer, but needed more time to evaluate a deal before making a formal bid. OpenRouter's founders did not want to wait, and Stripe ultimately prevailed with an $8 billion bid; Databricks and Stripe had already been in talks with the startup, and Nvidia's interest had not previously been reported. Over the following two months Nvidia signed more than $140 billion worth of deals, including a $105 billion credit guarantee, and held nearly $100 billion in equity investments plus $25 billion in future commitments as of late July.
An analysis argues that OpenRouter’s rankings measure token volume routed through one intermediary rather than global developer preference, and that prices, discounts, free access, batch jobs and self-testing can heavily influence the results. It also questions the comparability of Artificial Analysis scores and ARC-AGI results when test composition, weighting, scoring anchors, interfaces and versions change.
In a set of reflections from a week in San Francisco, Ben's Bites sorts AI users into roughly four groups: everyday people treating it as a better Google, curious non-engineers who try Lovable or Replit but quit at the first database bug, non-engineers who can steer a coding agent, and professional developers. The author argues personal agents will be the interface of the future but aren't yet, citing three pet peeves: one thread for everything, invisible memory, and loops that never close across apps. The piece also notes briefly that Claude Code can now be modded like a video game and that Google Docs opens Markdown files directly.
An analysis of two Andreessen Horowitz reports argues that AI adoption is broader than its paid and intensive use would suggest. It highlights a widening gap between casual users and power users, falling token costs driving greater demand for agents, and AI infrastructure spending flowing into chips, energy, and other physical assets.
What problem it solves
AI agents need to choose among models while balancing task success, cost, and latency for their workloads. Applications also need flexibility in where routing runs and who controls model clients, credentials, retries, and service operation.
Key features
Multiple routing algorithms — Routing can be configured with Auto, Task, Execution, or Composite strategies, including mode…
What problem it solves
Researchers, red-teamers, and system tinkerers need a chat interface that can work with multiple hosted and local models while exposing evaluation, perturbation, and sampling controls. They also need clear visibility into data flows, storage, telemetry, and deployment boundaries so they can make informed choices about privacy and self-hosting.
Key features
Multi-model races …
What problem it solves
Useful software engineering agents can require complex control flows, specialized tools, extensive configuration, and large dependency footprints. This project provides a simple, hackable agent that can work through the shell and run in local or sandboxed environments.
Key features
Shell-only agent — The agent interacts with its environment through bash without requiring mod…
What problem it solves
Teams that rely on AI agents need a way to give each agent a role, workspace, and instructions, keep context across conversations, and control what agents are allowed to do. Without that, agents run blind: actions happen without review, memory does not persist, and repeating scheduled work means redoing setup by hand.
Key features
Create agents and teams — Give an agent a na…