openrouter.ai

AIAI-generated summary OpenRouter reports that LLM quantization stores weights in fewer bits to reduce memory and hardware requirements. OpenRouter tested 1,455 endpoints across 470 models on October 20, 2026, to measure the effects of precisions like fp8 and fp4.

AIAI-generated summary OpenRouter Blog reports that giving an agent a Model Context Protocol server adds tens of thousands of tokens to the context window on every request. OpenRouter Blog tested tools including a command-line tool, a skill, and the GitHub MCP server on models such as Claude Sonnet 5.5 and GPT-6 Luna.

AIAI-generated summary OpenRouter explains how API rate limits and 429 errors occur when using OpenAI, Anthropic, and OpenRouter. OpenRouter details fixes including waiting for reset times, using fallback models, and routing requests through different providers.

AIAI-generated summary OpenRouter Blog published a tutorial on building a voice AI restaurant booking agent in Python that transcribes callers, uses tools to check availability, and converts responses into speech. OpenRouter Blog states that across 12 test turns, the median server-side time to first audio was 1.9 seconds, with text-to-speech accounting for roughly 90% of the measured model cost.

AIAI-generated summary OpenRouter Blog tested 15 vision models, Mistral OCR, and Tesseract on eight custom test pages to measure text extraction performance. OpenRouter Blog reports that Claude Opus 5.5 achieved the most correct key values, while Mistral OCR delivered the best value.

AIAI-generated summary OpenRouter explains how to obtain reliable structured output from any large language model using JSON schemas, tool calls, or decision models. OpenRouter says users can configure strict schemas, use response healing, and validate every reply to prevent parsing failures and code breaks.

AIAI-generated summary OpenRouter says it is launching a new generation of Auto Router and a specialized Auto Code router on OpenRouter. OpenRouter says Auto Code scores in line with Claude Opus 5.5 across five benchmarks while costing 74% less per task.

AIAI-generated summary OpenRouter available data shows open-weight models grew to 72% of top 50 token volume in late September 2026. OpenRouter states the token-weighted list price for open-weight models was $0.28 per million input tokens, while closed-weight models cost about four times as much.

AIAI-generated summary OpenRouter launched a new provider routing algorithm designed to reward providers for delivering fast, high-throughput, and reliable inference. OpenRouter says the algorithm has unlocked improvements in time to first token and throughput, along with a reduction in capacity errors.