Blog
OpenRouter shares news, updates, and posts.
openrouter.aiCrawled Oct 15, 2026
OpenRouter shares news, updates, and posts.
OpenRouter builds tools to help developers ship AI products.
OpenRouter offers Free, Standard, Business, and Enterprise plans, with access to 500+ AI models and payment based on usage.
Compare AI models by key metrics including benchmarks, price, context length, and other model features. View detailed stats for any model such as latency, uptime, and throughput.
Compare Qwen3.7 Flash from Qwen and Qwen3.7 Max from Qwen on key metrics including benchmarks, price, context length, and other model features.
Compare Qwen3.7 Flash from Qwen and Qwen3.7 Plus from Qwen on key metrics including benchmarks, price, context length, and other model features.
Compare Qwen3.8 2.4T A95B from Qwen and Qwen3.8 Max (0902) from Qwen on key metrics including benchmarks, price, context length, and other model features.
Compare Qwen3.8 27B from Qwen and Qwen3.8 2.4T A95B from Qwen on key metrics including benchmarks, price, context length, and other model features.
Compare Qwen3.8 27B from Qwen and Qwen3.8 Max (0902) from Qwen on key metrics including benchmarks, price, context length, and other model features.
OpenRouter provides developers with seamless access to 500+ AI models through a single API. Backed by leading investors and trusted by millions.
The page describes a unified interface for models, with model and price information for prompts.
The page explains how OpenRouter collects, uses, discloses, and protects personal data.
End User License Agreement for OpenRouter Stealth Program models, covering access, use, data training, and acceptable use policies.
Supplemental terms for individual Stealth Program models on OpenRouter, as required by their respective providers.
The page outlines the terms and conditions governing use of OpenRouter and its services.
OpenAI released its **Decisions API**, a fast, cost-efficient alternative to LLMs, to all users, competing with TypeSafe’s Jev model. The API, derived from ChatGPT Luna, reduces inference time to **100–180 milliseconds** and is already adopted by companies like **Box, Datadog, and Speak** for high-throughput tasks. Early tests show **Speak cut response times from 1.4 to 0.13 seconds**, while OpenAI improved its **Dots voice assistant** with the model. Benchmarks favor **Liquid’s D1 Ultra and Base Lab’s Elephant-1**, but Jev leads in adoption and excels in **PacMan-style decision-making tests**. OpenAI’s model stands out for **multi-modal capabilities**, including image classification, though speed optimizations could unlock real-time applications like **per-frame video processing**.
Inception AI Inc. introduces **Mercury 3**, a diffusion-based language model that outperforms OpenAI’s **GPT-6 Luna** and Anthropic’s **Claude Haiku 5.5** in speed, achieving up to **8x faster token generation** (1,068 tokens/sec vs. ~240 for Haiku 5.5). The model excels in coding tasks like **SWE-Bench Atlas** (51.1 vs. 33.3 for GPT-6) but trails in long-context tests. Early adopters—such as **Octen AI** (real-time search) and **RevisionDojo** (tutoring)—report sub-second response times, addressing latency-sensitive use cases. Pricing starts at **$0.20M input/$0.70M output tokens**, with API compatibility via OpenRouter, Baseten, and Vercel. Founded in 2024, Inception raised **$50M** (led by Menlo Ventures) and was reportedly in **acquisition talks with Microsoft** (valuation >$1B).
Chinese artificial intelligence developer MiniMax Group Inc. today released M3.1 Flash, a faster and cheaper update to the M3 large language model it launched in June. The release takes aim at the cost of running AI agents, which often need dozens of model calls to finish a single job. MiniMax said
Reflection AI released **Beam**, a 501-billion-parameter open model, designed to complete tasks using fewer tokens than comparable models, reducing per-task costs. The company trained Beam with reinforcement learning, penalizing unnecessary token usage while ensuring task completion. Beam outperforms competitors like DeepSeek’s V4.1 Flash in token efficiency, according to benchmarking firm Artificial Analysis. While Chinese models dominate performance rankings, U.S. firms like Pinterest, DoorDash, and Coinbase have adopted them for cost savings. Reflection positions Beam as a lean alternative, targeting enterprise use cases like coding and agent workflows.
AI agent startup Decagon held early merger-and-acquisition talks with ServiceNow and other parties. The company could be the latest startup swept up in a growing wave of AI dealmaking.
Nvidia caught wind in July of a looming takeover of OpenRouter, a three-year-old startup that had become a popular marketplace for AI models.
OpenAI gains ground on Anthropic as competition over AI pricing intensifies. Both companies need to demonstrate sustainable businesses to Wall Street ahead of planned IPOs, while Anthropic has edged to the front.
Peterson, Cisco Systems Inc.’s vice president of product management for Secure WAN, argues that enterprise AI needs dynamic model routing grounded in context, policy, visibility and controls. He compares model routing to SD-WAN, which steers traffic using application and network signals, and says model selection should account for task requirements, performance, cost, privacy and current conditions. AI workflows spanning branches, clouds, SaaS, data stores and services make adaptive routing especially important at the edge.
Developers are using Claude Code’s harness to run non-Anthropic models, including OpenAI’s GPT-5.6 Sol, through proxies and services such as OpenRouter. The approach lets them use the coding tool with models described as cheaper alternatives to Anthropic’s, rather than relying exclusively on Anthropic models.
OpenRouter says Stripe has agreed to acquire the company. The transaction remains subject to customary closing conditions and is expected to close in the coming weeks.
Stripe reportedly finalizes a deal to acquire OpenRouter at a price above $7 billion, according to Bloomberg. The AI gateway lets customers select among models based on their needs and budget; OpenRouter said it had 8 million global users and access to more than 400 models, after raising $113 million in a Series B at a reported $1.3 billion valuation in May.
Stripe is weighing a roughly $10 billion acquisition of OpenRouter, an AI model marketplace, as it pushes further into the AI economy. The potential purchase is presented as a strategic bet on that market, but the supplied text gives no details about OpenRouter’s products, customers, or deal status beyond the proposed acquisition.
Stripe is weighing an approximately $10 billion acquisition of AI model marketplace OpenRouter. The proposed price would be a steep increase from OpenRouter’s $1.3 billion valuation as Stripe extends into AI infrastructure and model access.
OpenRouter reaches a run rate of about 1.5 quadrillion tokens per year and has more than 8 million developers. The company raises a $113 million Series B at a $1.3 billion valuation, with CapitalG leading the round.
OpenRouter raises $113 million in a Series B led by CapitalG. Existing investors Andreessen Horowitz, Menlo Ventures, and NVIDIA's NVentures also participate, as weekly token volume reaches 25 trillion.
Menlo Ventures invests $40 million in OpenRouter's Series A, describing the company as a single API for accessing AI models. OpenRouter gives more than 2.5 million developers access to over 400 models from more than 60 providers, with centralized billing.

OpenRouter is a unified API and marketplace for 500+ AI models from 80+ providers.