# Brainbase > Brainbase hosts your agents, scores every run, and evaluates every change before it ships — across any model or harness. Every page on this site has a Markdown mirror: append `.md` to the path (`https://brainbaselabs.com/pricing.md`), or `/index.md` for the home page. The mirrors are generated from the pages themselves, so they never disagree with what a person reads. They carry no images, navigation, or scripts. ## Pages - [Home](https://brainbaselabs.com/index.md): what Brainbase is — hosted agents, evals, and the self-improvement loop. - [Pricing](https://brainbaselabs.com/pricing.md): plans and what each includes. - [Company](https://brainbaselabs.com/company.md): who is building this, and why. - [Contact sales](https://brainbaselabs.com/contact.md): how to reach the team. - [Blog](https://brainbaselabs.com/blog.md): product updates and technical writing. ## Posts - [Event recap: Startup Speedrun Hackathon](https://brainbaselabs.com/blog/startup-speedrun-hackathon-recap/index.md): Brainbase, Anthropic, Cloudflare and Stripe gave builders at Cloudflare HQ one day to take an idea as far as it would go. Twenty-nine projects shipped. Here are the top five. - [Panel recap: How to build for agents that act in the real world?](https://brainbaselabs.com/blog/panel-agents-that-act-in-the-real-world/index.md): Brainbase, Taste Labs, Stripe and Cloudflare on taste, payments, infrastructure and internal agents: what changes when agents stop answering and start acting. - [@anthropic/claude-opus-5.5 is live on Brainbase](https://brainbaselabs.com/blog/anthropic-claude-opus-5-5-live-on-brainbase/index.md): Claude Opus 5.5 performs at Fable 5.1 level on most work, leads on agentic coding, and costs about 40% less to run than Opus 5. It is on every Brainbase harness today. - [@openai/gpt-6-sol and @openai/gpt-6-luna are live on Brainbase](https://brainbaselabs.com/blog/openai-gpt-6-sol-and-luna-live-on-brainbase/index.md): OpenAI's two cheaper GPT-6 models: Sol at $2 / $10 for everyday coding and agent work, Luna at $0.10 / $0.50 for high-volume tasks, both at half the price of their GPT-5.6 predecessors and both on every Brainbase harness today. - [@spacexai/grok-4.7 is live on Brainbase](https://brainbaselabs.com/blog/spacexai-grok-4-7-live-on-brainbase/index.md): Grok 4.7 is a larger base model with a longer reinforcement learning run, better at checking its own work and at multi-hour terminal and office tasks, at the same price and speed as Grok 4.6. Now on every Brainbase harness. - [Automating Domino's local competitor pricing research](https://brainbaselabs.com/blog/dominos-competitor-pricing/index.md): How the Brainbase team embedded with Domino's CMO office to build a competitor pricing agent from the ground up, and turned a judgment-heavy research task into a measurable loop: accuracy from 41% to 97%, cost per source down 52%, and 41 minutes back per run. - [@openai/gpt-6-astra is live on Brainbase](https://brainbaselabs.com/blog/openai-gpt-6-astra-live-on-brainbase/index.md): GPT-6 Astra is OpenAI's first true long-horizon agent: it leads on terminal, repository, and computer-use benchmarks, works software through the screen, and costs 2.5× its predecessor. - [@meta/muse-spark-1.3 is live on Brainbase](https://brainbaselabs.com/blog/meta-muse-spark-1-3-live-on-brainbase/index.md): Muse Spark 1.3 closes most of the gap to Opus 5 and GPT-5.6 Sol on agentic coding, leads outright on long context, and does it with 20% fewer tool calls at unchanged prices. - [@google/gemini-3.8-flash is live on Brainbase](https://brainbaselabs.com/blog/google-gemini-3-8-flash-live-on-brainbase/index.md): Gemini 3.8 Flash puts Opus-class scores on finance, legal, and science agents at a fraction of the price — with a real gap left on long-horizon terminal work. - [@anthropic/claude-fable-5.1 is live on Brainbase](https://brainbaselabs.com/blog/anthropic-claude-fable-5-1-live-on-brainbase/index.md): Claude Fable 5.1 more than doubles its predecessor on agentic science, adds a five-step effort dial, and cuts cache reads by 75% — available on every Brainbase harness today. - [@zai/glm-5.3-flash is live on Brainbase](https://brainbaselabs.com/blog/zai-glm-5-3-flash-live-on-brainbase/index.md): GLM-5.3-Flash — the model behind the Ox Alpha stealth preview — brings native multimodality, a 1M-token context window, and MIT-licensed open weights to every Brainbase harness. - [Introducing Universal Managed Agents API](https://brainbaselabs.com/blog/universal-managed-agents/index.md): One API to run agents across eight harnesses and any model, in Modal, Daytona, or E2B sandboxes — configured at creation with the open Universal Agent Protocol. - [@spacexai/grok-4.6 is live on Brainbase](https://brainbaselabs.com/blog/spacexai-grok-4-6-live-on-brainbase/index.md): Grok 4.6 brings frontier agentic performance, a 500K context window, and aggressive token pricing to every Brainbase harness. - [@meta/muse-spark is live on Brainbase](https://brainbaselabs.com/blog/meta-muse-spark-live-on-brainbase/index.md): Muse Spark 1.2 brings Meta's coding-optimized agent model, 1M tokens of context, and efficient API pricing to Brainbase. - [@thinking-machines/inkling is live on Brainbase](https://brainbaselabs.com/blog/thinking-machines-inkling-live-on-brainbase/index.md): Brainbase now hosts Inkling, the newest open-weights model from Thinking Machines, so teams can deploy and test Inkling AI agents on day zero. - [Brainbase is now generally available](https://brainbaselabs.com/blog/brainbase-is-now-ga/index.md): Introducing Brainbase, the first AI agent cloud and our first step toward a future of non-monolithic intelligence that provisions like compute. ## Elsewhere - [Documentation](https://docs.brainbaselabs.com) - [API reference](https://docs.brainbaselabs.com/api) - [Dashboard](https://app.brainbaselabs.com/) ## Legal - [Terms of Service](https://brainbaselabs.com/terms.md) - [Privacy Policy](https://brainbaselabs.com/privacy.md)