正在加载视频...

视频加载失败

today, we're launching Light Parse Extend can now parse documents up to 70% cheaper than other solutions - light parse starts at $0.00625 per page, and scales down significantly with volume discounts - it's SOTA on Databricks OfficeQA Pro (a 3rd party agent benchmark) among all non-agentic parsing providers...

58,356 次观看 • 2 个月前 •via X (Twitter)

45 条评论

Kushal Byatnal 的头像
Kushal Byatnal2 个月前

light parse outperforms and outprices alternatives on document parsing Databricks OfficeQA Pro is a benchmark that measures how accurately an LLM can answer questions from scanned U.S. government financial documents this benchmark is truly independent. we ran the evaluation against it, but we had no part in designing it (nor did we commission it)

Ammaar Reshi 的头像
Ammaar Reshi2 个月前

whenever I see kb in a video, I’m airplaying it to my TV

Kushal Byatnal 的头像
Kushal Byatnal2 个月前

filming this video is what convinced me to go get a haircut

Ammaar Reshi 的头像
Ammaar Reshi2 个月前

I don’t see the difference

Davis Treybig 的头像
Davis Treybig2 个月前

Extend defining the full Pareto frontier for document processing!

Kushal Byatnal 的头像
Kushal Byatnal2 个月前

🫡

Pri 的头像
Pri2 个月前

Pleasure working on the launch video! 🔥

aiden brotman 的头像
aiden brotman2 个月前

even more affordable and accurate? what can't extend do?

Darren Chow 的头像
Darren Chow2 个月前

@chaneratextend extend ml team is so cracked

Kushal Byatnal 的头像
Kushal Byatnal2 个月前

@chaneratextend wait until they see next weeks launch

bo lau 的头像
bo lau2 个月前

this is huge

Kushal Byatnal 的头像
Kushal Byatnal2 个月前

don't tell them what's coming next week

scott 的头像
scott2 个月前

mr krabs approves

Hansen Qian 的头像
Hansen Qian2 个月前

Congrats!!! Big accomplishment

Kushal Byatnal 的头像
Kushal Byatnal2 个月前

ty hansen!

Mike Bedarev 的头像
Mike Bedarev2 个月前

The Pareto Frontier had no idea what our next move was

Victor Dubeux 的头像
Victor Dubeux2 个月前

light & cheap, let's go!!!!

James Russo 的头像
James Russo2 个月前

Congrats on the launch!

Kushal Byatnal 的头像
Kushal Byatnal2 个月前

thanks James!

Jordan Meyer 的头像
Jordan Meyer2 个月前

they said it couldn't be done

Kushal Byatnal 的头像
Kushal Byatnal2 个月前

that's because they've never met @ameyaajoshi and @joon_kim96

Cody 的头像
Cody2 个月前

ACORD forms?

Kushal Byatnal 的头像
Kushal Byatnal2 个月前

yup! very common use case for us

Pierre-Louis Vouteau 的头像
Pierre-Louis Vouteau2 个月前

@ycombinator Congrats!!

Kushal Byatnal 的头像
Kushal Byatnal2 个月前

@ycombinator thank you! we are big fans of @Ooakdata btw 🔥

Guhan 的头像
Guhan2 个月前

Congrats to your team. Cost really matters; I am currently using LlamaParse. Would you be able to explain how your product works? Is it agentic because I am working on legal workflows that need pure reflection rather than interpretation?

John Wei 的头像
John Wei2 个月前

congrats, lfg!

Arvind 的头像
Arvind2 个月前

🔥 always cooking

Kushal Byatnal 的头像
Kushal Byatnal2 个月前

a true chef never stops

Gordon Wintrob 的头像
Gordon Wintrob2 个月前

Congrats Kushal and team!! ⚡️⚡️⚡️

Launch Archive 的头像
Launch Archive2 个月前

Light Parse offers a cost-effective solution for document parsing, making it easier for developers to manage large data. Congrats @kushalbyatnal on #2 on Launch Archive today.

Kalyan 的头像
Kalyan2 个月前

cheap parsing is huge - I lean on vision models to read grocery receipts in my new iOS pantry app:

Elise Tech Ai 的头像
Elise Tech Ai2 个月前

70% lower costs with strong benchmark performance is a huge win—efficient document parsing is getting seriously competitive.

RAZA | AI EXPLORER 的头像
RAZA | AI EXPLORER2 个月前

70% cheaper while maintaining strong benchmark performance is a huge win for document AI.

Luke Beck 的头像
Luke Beck2 个月前

How is strikethrough detection?

SomacoSF 的头像
SomacoSF2 个月前

add municipal permits by zipcode. meaning oull all munucipal permit data by zipcode so you can see what is being permitted for build. then you can extrapolate from there

Alice The Ai Expert 的头像
Alice The Ai Expert2 个月前

Huge win 70% cheaper with SOTA accuracy is exactly what teams need!

Lokesh Sanapalli 的头像
Lokesh Sanapalli2 个月前

Congratulations! How does it compare with firecrawl pdf inspector?

Internet Labs 的头像
Internet Labs2 个月前

At $0.00625 a page, I’m about to parse every obnoxiously long CVS receipt I’ve ever received just to test the table models. Huge win for the builders.

Saurabh Jambure 的头像
Saurabh Jambure2 个月前

Interesting

Romero Rodrigues 的头像
Romero Rodrigues2 个月前

70 percent cheaper on the page is the actual product. Doc AI just became a cost line, not a seat. Whoever owns the parse at a fraction of a cent owns the agent workflow that sits on top of it.

Haseeb 的头像
Haseeb2 个月前

70% cheaper gets especially interesting once parsing becomes a default step in every document workflow.

Modelplane 的头像
Modelplane2 个月前

The 70% cheaper claim is bold—curious how it holds up on complex layouts like tables or handwriting, where parsing costs usually spike. Databricks OfficeQA Pro is a solid benchmark, but real-world docs always find the edge cases.

Dinesh | Umbra 的头像
Dinesh | Umbra2 个月前

the real ai race is becoming a race to zero inference cost

Vakada Rohit 的头像
Vakada Rohit2 个月前

70% cheaper without sacrificing accuracy is a seriously impressive improvement.

相关视频

Today we’re releasing OCRia-1, a frontier VLM for document parsing that sets a new SOTA in accuracy, cost, and latency. It not only sets a new high score on ParseBench (87.4), but also leads in three 3rd-party benchmarks: olmOCR-bench* (83.5 overall, 92.9 on tables!), PulseBench-Tab (84.1), and fr-bench-pdf2md (72.1). OCRia-1 is really, really good at parsing complex tables: nested headers, merged cells, irregular grids, scanned tables, and more. It is also fantastic at parsing complex unlabeled charts (line, bar, pie, technical diagrams), and preserving complex formatting. Most real-world enterprise documents contain at least one of the these, whether it’s a big tax withholding table or a public economic report. It is a 4B model and outperforms all other models within a similar parameter class: LightOnOCR-3, Chandra-OCR 2, Mistral OCR 4.1, and more. It is also extremely fast. All credit goes to our research team, who spent countless hours on curating a diverse dataset, tuning our training pipeline, and squeezing every last drop out of model optimization so that it can process billions of pages at scale. OCRia-1 as a base is powering all of our modes within LlamaParse (cost-effective, agentic, agentic plus), so you can access its capabilities today! If you’re processing millions/billions of documents at scale, we’re offering enterprises that book a demo up to $3k in credits for a guided evaluation.** *olmOCR-bench scores are adjusted based on unifying rendering at 1,288px for all models **enterprise POC credits apply to eligible organizations that book a demo

Jerry Liu

111,973 次观看 • 25 天前