Narev AI

Doing boring bits of FinOps and RevOps for AI

[Narev AI @narev_ai]

@alphatozeta8148 @vllm_project https://t.co/W4gzDuq7KN

[Narev AI @narev_ai]

Clerk did for auth what @ai-billing does for AI billing.

Wrap your model, define a destination, and every request is automatically metered and charged. No billing infra to build.

[Narev AI @narev_ai]

You can track the price cuts with our real-time pricing MCP.

https://t.co/3cMkmu8tOD

[Polymarket @Polymarket]

NEW: OpenAI is reportedly considering drastic price cuts as it anticipates a “war” for users with Anthropic.

[Narev AI @narev_ai]

This is what these charts look like if you have axes start from 0

[Cursor @cursor_ai]

Claude Fable 5 is now available in Cursor.

It sets a new state of the art on CursorBench at 72.9%, 8 points above the previous best.

[Narev AI @narev_ai]

@JTBeers @cursor_ai So https://t.co/85ipsMnIAi actually has the axes other way around. Is it any clearer?

[Narev AI @narev_ai]

Adding usage-based billing to your AI app shouldn't take a weekend.

@ai-billing wraps the Vercel AI SDK - one middleware, 5 lines of code.

npm install @ai-billing/core and you're done.

https://t.co/RsAGPWFCCm

[Narev AI @narev_ai]

https://t.co/RsAGPWFCCm

[Narev AI @narev_ai]

Good job @stripe with @𝘀𝘁𝗿𝗶𝗽𝗲/𝗮𝗶-𝘀𝗱𝗸 package. Users of Lago, Openmeter and Polar - we got you too with @𝗮𝗶-𝗯𝗶𝗹𝗹𝗶𝗻𝗴/𝗹𝗮𝗴𝗼 @𝗮𝗶-𝗯𝗶𝗹𝗹𝗶𝗻𝗴/𝗼𝗽𝗲𝗻𝗺𝗲𝘁𝗲𝗿 and @𝗮𝗶-𝗯𝗶𝗹𝗹𝗶𝗻𝗴/𝗹𝗮𝗴𝗼.

[Narev AI @narev_ai]

If you want to implement usage-based billing, just like @GitHubCopilot did today, we can help.

https://t.co/v21OMAZtnK

[Narev AI @narev_ai]

@poolsideai @poolsideai How can we track the price of Laguna models in real time? Do you have an endpoint we can ping?

[Narev AI @narev_ai]

2/2 OpenRouter pricing:

The raw pricing numbers side-by-side per 1M tokens for grok-4.20-multi-agent-0309:
@xai Direct:
Input: $1.25
Output: $2.50

@OpenRouter:
Input: $2.00 (≤200k). $4.00 (>200k)
Output: $6.00 (≤200k) ➔ $12.00 (>200k)

https://t.co/S6WL3uno1r
https://t.co/9fAQtP7pz3

[Narev AI @narev_ai]

1/2 Using LLM providers directly saves budgets.

You're paying 1.6x more for input tokens and 2.4x more for output tokens using Grok 4.20 Multi-Agent at @OpenRouter vs going directly to @xai

And if you cross the 200k context window? That premium jumps to over 3x and 4x.

[Narev AI @narev_ai]

Team Narev shipped beautiful billing components to narevai/ai-billing.

Adding usage-based billing to a Next.js app is now as simple as adding two lines <CreditTopUpPolar/> <CreditUsagePolar/>

[Narev AI @narev_ai]

Release 0.2.6

https://t.co/y2RlZ9HOoH

[Narev AI @narev_ai]

Pure spring cleaning in our open source project. We've overhauled the dependency stack to the latest lines (node v24, hello).

Hats off to the @Mend_io team for their Github bot. So easy to get everything updated.

[Narev AI @narev_ai]

@OpenRouter are we going to see DeepInfra added as a routing fallback for MiMo soon?

[Narev AI @narev_ai]

Did @DeepInfra just silently add Xiaomi MiMo models?

Pulled their models endpoint and spotted both V2.5 and V2.5-Pro sitting there.

Same pricing as Xiaomi's <=256K tier.

[Narev AI @narev_ai]

Looks like @DeepInfra are matching Xiaomi’s base tier direct pricing exactly:
- XiaomiMiMo/MiMo-V2.5-Pro: $1 In / $3 Out / $0.20 Cache Read
- XiaomiMiMo/MiMo-V2.5: $0.40 In / $2 Out / $0.08 Cache Read