dari.dev (YC F25)

the routing layer for ai agents

[Avyay Varadarajan @avyvar]

They spent millions on router dot com.

We spent $20 on https://t.co/dgHRbe0jEh.

We saved money on the domain to save you money on tokens.

> Pool Claude + Codex subs
> Pick the right model + thinking level w/ our SOTA router
> Use your traces to fix token waste
> Open-source

[Avyay Varadarajan @avyvar]

DeepSeek-V4.1-Flash-1017 is Pareto-mogging in our initial tests.

It's the first open-source model we've tested that's consistently on the Pareto frontier in real-world PRs, running in @pidotdev

Some results from Sentry, PostHog, and Next.js PRs:

[Avyay Varadarajan @avyvar]

Haiku 5.5 is an incredibly capable model at its size.

We ran Haiku on real Pi, Next.js, PostHog, and Sentry tasks, and it was on the Pareto frontier in every test.

Slightly higher $/task than Luna for increased performance.

Towards intelligence too cheap to meter!

[Avyay Varadarajan @avyvar]

Is @MistralAI Large 4 smart or just chonk-y?

We ran Le Chonk on real PRs from Next.js, Sentry, Pi, and PostHog.

> Mistral Large 4 generally benchmarked worse + similar price as Kimi K3 and GLM 5.3 in the Pi harness.
> None of the 3 models were ever on the Pareto frontier.

[Avyay Varadarajan @avyvar]

On our eval of 82 tasks from real @sentry PRs, GLM 5.3 was
> 50% more expensive than claude opus 5.5 ($1.83 vs $1.22 per task)
> 8.5 points less accurate (74.4% vs 82.9%)

[Avyay Varadarajan @avyvar]

Open-weight models may be costing you more on real work.

We built self-bench to turn real GitHub PRs into evals.

Testing repos like Next.js, PostHog, and Sentry, open models were usually pricier + less capable.

Find the best model for your repo: https://t.co/VhcZR2zsyq

[Avyay Varadarajan @avyvar]

if you're running out of Fable weekly limits try running

`dari --claude`

Use Fable + open models and make your subscriptions last for 2x longer with no quality degredation :)

[Avyay Varadarajan @avyvar]

Public evals are saturated. What matters is that a model reliably works in your codebase.

To solve this, we built self-bench (https://t.co/zYfPrbem2P) - an open source package to automatically build evals on your own private PRs.

One CLI command -> get a clean Harbor dataset.

[Chubby♨️ @kimmonismus]

79.8% on Terminal-Bench 2.1 for $76 in total inference cost.

That's the number Dari @daridotdev is launching with today. They open-sourced the router model behind it: a small fine-tuned model that decides which model handles each step, sending most of them to cheaper options and only calling the expensive frontier models when they actually change the result. It factors the cache into that decision too, so it won't switch models when the cached path is the cheaper win. On the same 89-task suite, the frontier setups run into the thousands.

Runs inside Claude Code, Codex, and Pi, you can bring your own Anthropic and OpenAI subscriptions, and the weights are on Hugging Face.

[Avyay Varadarajan @avyvar]

Today, we're releasing our open-weight, auto-routing model @daridotdev, built for coding agents.

We're state-of-the-art on the Pareto Frontier, w/ 70% cost reduction + comparable coding performance to Fable.

Bring your own evals, choose your models, or use our defaults.

[Avyay Varadarajan @avyvar]

Token-maxxing is getting out of hand.

Most AI apps send every request to the biggest model, even when a smaller model would work.

We built Dari Router to fix that.

[Avyay Varadarajan @avyvar]

We made deploying agents to prod so easy your coding agent can do it in under 5 minutes.

Get a production-grade API that scales to 100s of concurrent sessions, pauses for days between messages, and recovers sessions even on sandbox crashes or LLM provider failure.

@daridotdev

[Avyay Varadarajan @avyvar]

https://t.co/JA5P7ppdGg

[Avyay Varadarajan @avyvar]

Introducing @daridotdev.

Deploy your AI agent as a stateful, versioned API.

- durable sessions
- built-in execution + sandboxes
- tools, files, retries handled

Stop stitching infra. Just dari deploy.