OpenRouter

OpenAI slashed prices for its GPT-5.6 Luna and Terra models, but despite the price cuts, usage dramatically increased. This led to significant revenue gains, with Luna's revenue increasing about 34% and Terra's revenue rising about 45%. This pattern suggests an instance of Jevons Paradox, where making a resource cheaper encourages much higher consumption.

Companies are facing unexpected financial costs from running autonomous AI coding agents due to high token usage, resulting in sticker shock. Consequently, there has been a surge of demand for AI model routers—software that optimizes requests by selecting the best model for cost and performance. This technology allows organizations to potentially reduce inference costs by double-digit percentages.

A potential US ban on Chinese open-weight AI models could cost American businesses up to US$12 billion annually, according to a US-based academic. Usage data from OpenRouter, an LLM aggregator, suggests reliance on cost-efficient Chinese solutions is growing.

The article introduces world-model-optimizer, an open-source tool for distilling and serving smaller AI models that maintain frontier quality while reducing costs. It leverages agent traces to optimize models for efficiency and offers a hosted solution with over 40% cost savings.

Mira Murati's Inkling AI model, released after two years of silence from Thinking Machines Lab, is now available on OpenRouter. The model's MCP score is described as genuinely impressive, though its price-to-performance ratio is more complex.

China's open-weight AI models from companies like Moonshot AI, Tencent, and Xiaomi are gaining dominance over Silicon Valley's premium models, with Chinese models occupying the top five spots on OpenRouter by weekly token usage. Businesses are shifting to cheaper, customizable alternatives for routine tasks, threatening the market position of U.S. AI leaders like OpenAI and Anthropic.

OpenRouter's Fusion compound-model API combines budget AI models to outperform GPT-5.5 and Claude Opus 4.8 in benchmark tests. The release coincides with Fable 5's deactivation.

The U.S. government's ban on Anthropic's Fable and Mythos models for non-U.S. users has accelerated adoption of open-source AI, particularly from Chinese labs like DeepSeek and Moonshot AI. Chinese models, including Knowledge Atlas's GLM-5.2, are gaining traction globally due to their cost-effectiveness and accessibility, with four Chinese models now dominating OpenRouter's most-used list.

Startups like OpenRouter and Concentrate AI are gaining traction by helping companies manage AI model costs through routing tools, as cheaper models like DeepSeek's V4 emerge. These companies face competition from Big Tech firms like AWS, Microsoft, and Google Cloud.

OpenRouter secured $113 million in Series B funding. The article received 96 points and 26 comments on Hacker News.

OpenRouter has increased its valuation to $1.3 billion in a year, raising $113 million in a Series B funding round led by CapitalG. The company reported a 5x growth in usage over six months, signaling the rise of a multi-AI-model future.

Apple Silicon is compared to OpenRouter in terms of cost, with Apple's product being more expensive. The article discusses this price difference and its implications. The comparison was shared on Twitter by Rohan Sood.

Apple Silicon is compared to OpenRouter in terms of cost, with Apple Silicon being more expensive. The article discusses the energy use of offline LLM. The comparison is discussed on a blog post and commented on a news site.

Google reimbursed two developers who were victims of API fraud, but plans to continue automatically expanding users' spending limits, leaving them vulnerable to unexpected bills. One of the developers, Isuru Fonseka, had his $17,000 bill reversed after a hacker took control of his account. Despite the refund, Google seems to have lost Fonseka as a customer.

OpenAI's latest GPT-5.5 model is priced higher than its predecessor, with costs increasing by 49 to 92 percent despite token processing efficiency improvements. The price hike affects both short and long prompts, with longer prompts seeing more offset by shorter completions. OpenAI faces significant projected losses, with a reported $14 billion loss in 2026.

DeepClaude allows users to run Claude code at a lower cost by swapping the expensive Anthropic backend with alternative AI options. This new open-source script reduces expenses while maintaining the agent loop. The cost is 17 times cheaper than the original option.

Eden AI is presented as a European alternative to OpenRouter, an AI model provider. The article, hosted on Eden AI's website, has gained traction on Hacker News with 58 points and 31 comments.

The article discusses reallocating $100/month spent on Claude code services to Zed and OpenRouter. This shift reflects a strategic adjustment in AI tool usage for cost efficiency and performance optimization.