InfraOpenRouterConfirmed
Story 3 of 11 in the Friday 6 November editionOpenRouter launches next-generation Auto Router and Auto Code for coding agents
The new routers pick a model and reasoning effort per request within workspace guardrails, matching Claude Opus 5.5 on coding benchmarks at 74% lower cost.

OpenRouter launched its next-generation Auto Router (openrouter/auto) alongside a dedicated coding router, Auto Code (openrouter/auto-code), on 6 November 2026. Processing more than 2.5T tokens per month, the updated routing engine evaluates each incoming prompt to select both the target model and the reasoning effort level from the pool permitted by the caller’s zero data retention (ZDR), provider, and bring-your-own-key policies.

Auto Code is tuned specifically for multi-turn software engineering and agentic tool use, achieving 74.3% on Deep SWE at $1.98 per task and cutting cost per task by roughly 74% compared with single frontier model calls at comparable accuracy. Both routers operate with a p90 routing decision latency of 0.65 seconds and expose per-request routing rationale and candidate model selection shares in the OpenRouter chat interface.

Sources
- PrimaryIntroducing the New Auto and Auto Code Routersopenrouter.ai · OpenRouter announcement
- On XOpenRouter next-generation routers postx.com · Official X announcement
The rest of the day
Read the whole day01ProductMeta
Meta launches Spaces in Muse for custom interactive tools built from chat
Users can describe an idea to build persistent planners, trackers, and games inside the Muse app, pin them to navigation, and keep them updated through connected apps.
02ModelMeta
Meta rolls out specialized Muse and Meta AI models with Active Memory
Trained using Muse Wonder, the upgraded models bring Active Memory and Active Focus to consumer Muse and Meta AI interfaces with 2.4 times lower critical forgetting.
04ResearchOpenAI
OpenAI and Ligo introduce BioBasicsBench for everyday wet-lab biology
The 100-task benchmark tests whether AI agents can interpret raw gels, flow cytometry plots, plates, and microscopy images, with GPT-6 Astra scoring 66.7%.
05ResearchAnthropic
Anthropic details Claude agent memory architecture and offline dreaming
An Engineering Blog post explains how background consolidation reviews transcripts, extracts corrections, and injects a compact memory layer into new sessions.
Spotted an error? Email aiforsuccess.in+compass@gmail.com or tag @ai_for_success on X







