# Extract, now 4x faster ⚡️ | LlamaIndex Newsletter

Canonical: https://brew.new/templates/llamaindex/extract-now-4x-faster-llamaindex-newsletter

Brand: llamaindex.ai
Category: newsletter

![Preview of Extract, now 4x faster ⚡️ | LlamaIndex Newsletter](https://cdn.brew.new/email-preview-181aa4bbb8f97cfe-tracking_r57g09trkzv7kr0ym39yr5j9718f1mcx-1790270475717.png)

## Email content

Email Newsletter 1200x360- H2-2026-2

Hey LlamaIndex Community! 🦙

We've been heads-down shipping, but there's a ton to catch you up on with new launches, a content series, and plenty of IRL time with the community. ➡️

⚡️ Extract, now 4x faster. ⚡️

Our fastest extraction yet: 4x faster than Cost Effective, at 3.7 seconds per page. Latency stays nearly flat as documents get longer, so it's built for your real-time uploads. Try it here!

🚢 JUST SHIPPED

LiteParse v2.14.6: Text-based PDFs now parse 25% faster. Still open source, and still local. LiteParse runs locally in Python, Node.js, Rust, or directly in the browser.

Spreadsheets, run natively: Text-based Native spreadsheet extraction is now built into LlamaParse! We read raw cells directly instead of flattening to text, so every value keeps its context.

We got on the Jev wave too: If you've been online this past week, you've probably seen people building all kinds of things with Jev. We gave it a try and built DocJev, which can classify and split docs 6x faster than gpt-5.6-luna with equivalent accuracy. You can pick between liteparse or LlamaParse for your OCR backend. Check it out!

👋🏽 IRL

RECENTLY:

U.S. Open: We hosted LlamaIndex customers and friends in an executive suite at the US Open in NYC. We're grateful to everyone who joined us to watch some great tennis!

Connected Stack: We joined Founder Flash Talks to talk about the struggles of enterprise agents dealing with messy, complex documents.

GMI Cloud Modern Renaissance: We had a great time speaking to CTOs and AI leaders about the agentic layer beneath multimodal systems.

COMING UP:

6

Insurance docs are brutal: scanned policies, claims files, tables that span pages. We're taking them on live and building a working pipeline from scratch with LlamaParse and Extract. Show up with your hardest doc questions! RSVP HERE →

Sep 28 - Founders Dinner 03: Evals & RL Envs (SF)

Sep 29 - Braintrust, LlamaIndex, Kernel: Agent Builders Night (NYC)

Sep 29 - Daytona, LlamaIndex, WorkOS: Builders Night (SF)

Oct 6 - Clickhouse, LlamaIndex, Artie: Tech Week AI Demo Night (SF)

Oct 9 - LlamaIndex, Lemma, TinyFish, MongoDB: Agent Economy Gala (SF)

🚨 SIGNAL

Anthropic and OpenAI cut prices the same day.

Claude Opus 5.5 launched at $4/$20 per million tokens, and 90 minutes later OpenAI shipped GPT-6 Sol at $2/$10 and Luna at $0.10/$0.50.

TypeSafe AI launched Jev, a "System One" model.

TypeSafe reports 200x faster inference and 400x lower cost than comparable LLMs on classification tasks.

Xiaomi open-sourced MiMo-V2.6 under MIT.

It ships as an omnimodal Pro model plus a 309B-parameter Flash MoE with 15B active parameters and 256K context.

Gradient_Newsletter

TRY FOR FREE

SEE PRICING

Building a new agent? Get in touch with our team.

Need support? Join our Discord.

Check out our Github for the latest code releases and documentation.

Thanks for reading this week's newsletter :)

We want your thoughts! Feel free to reply directly or email news@llamaindex.ai.

LlamaIndex-logo

LlamaIndex, 405 Howard Street, San Francisco, California 94105, United States

Unsubscribe

Manage preferences

github

discord

X

LinkedIn

YouTube

Outline_Text_Newsletter

[Open and remix this design](https://brew.new/templates/llamaindex/extract-now-4x-faster-llamaindex-newsletter)

[Browse email designs](https://brew.new/browse/templates)