Laya

What hackers pick instead

• Jev by Typesafe AIA cheap, low-latency general classifier you redefine with a prompt instead of retraining, though critics call it overhyped and little more than standard LLM tech.392recommend
• Cloudflare ClefOpen-weight decision model people like for its open-source license, self-hosting and image input; critics say it is slower, costlier and less accurate than Jev.21recommend
• OpenJevFast, cheap structured-output model that cannot emit tokens outside a given set. Critics doubt it differs from other small LLMs, and the naming is confusing.21recommend
• Typesafe AIPeople recommend it as a cheap API model that handles new classification problems without fine-tuning. Critics say it is over-hyped and not production-ready without fine-tuning.16recommend
• Strands Decider 2BA small open-source decision model that people run locally for filtering and automations. Critics note a bias toward the first option listed.7recommend

Every quote

The 40 most detailed of 85 comments. Each links to the original.

1. Having it open source is awesome as fine tuning might give additional performance on the task we care about.

   Recommends · Sep 2026 · in “I built non-autoregressive decision models with RL a year ago”

2. Laya is actually pretty cool. I think it will catch on considering its open weight and self hostable.

   Recommends · Sep 2026 · in “I built non-autoregressive decision models with RL a year ago”

3. After multiple iterations and improvements I have a version that is able to get results in 100-200ms.

   Recommends · Oct 2026

4. it was a drop-in replacement and it makes me happy I can get use out of old hardware I have running

   Recommends · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

5. Laya is designed to be fine-tuned for a specific task, and the site reports the fine-tuning gains.

   Mixed · Sep 2026 · in “Jeff – Jev-compatible 0.8B decision models, trained at home, ~30 ms”

6. saying "no churn risk" about a refund, in a leading example on their homepage, flabbergasting

   Mixed · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

7. you should use Laya for much more deterministic tasks where you have some training data

   Recommends · Sep 2026 · in “Laya on Mac M4 CoreML Offline”

8. Guess I'm about to become a big Laya fan if it can run on this kind of hardware

   Recommends · Sep 2026 · in “Laya on Mac M4 CoreML Offline”

9. the workflow is open source. The downside is that laya is only 1024 tokens

   Mixed · Sep 2026 · in “Jev – A curation of Jev demos on X, tools, skills, and integrations”

10. When I discovered Laya I realized we might have a feasible solution here.

    Recommends · Oct 2026

11. Laya was significantly more accurate (in this case) and almost as fast.

    Recommends · Oct 2026 · in “EmbeddingGemma 2: An open, lightweight multimodal embedding model”

12. fine tuning took 15m or so. Just wish it had a larger context window

    Mixed · Sep 2026 · in “Laya on Mac M4 CoreML Offline”

13. i can run simple tasks on local machine offline. its just cool.

    Recommends · Sep 2026 · in “Laya on Mac M4 CoreML Offline”

14. it will then turn on the lights and heater nearly instantly

    Recommends · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

15. great project for empowering open-source alternatives

    Recommends · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

16. I have mixed results with Laya on a pdf classifier

    Mixed · Oct 2026 · in “Strands Decider 2B: a small, open-source, decision model”

17. Laya was preferred due to need for low latency

    Recommends · Oct 2026 · in “Training a local model to play browser games on the fly”

18. Laya's out of the box model is much weaker.

    Mixed · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

19. Because what they did is kinda trivial.

    Mixed · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

20. I am able to get responses under 200ms

    Recommends · Oct 2026 · in “Build your own decision model”

21. you could try fine-tuning Laya models

    Mixed · Oct 2026 · in “Testing Jev, a fast multiple-choice AI model, on forensic accounting tasks”

22. No need for GPUs or tons of memory.

    Recommends · Oct 2026 · in “Typesafe AI raises $870M at $7.5B”

23. Set it up locally and it works!

    Recommends · Oct 2026 · in “Clef: Open-weight decision models, and new RL fine-tuning platform”

24. From my experience, Laya performs significantly worse. It's less confident and often makes wrong decisions with more complex queries.

    Criticizes · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

25. it is presented as groundbreaking, revolutionary, novel frontier tech when it is a known classification system

    Criticizes · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

26. Typesafe false marketing caused Laya to fit perfectly into pretty much every advantage that they are claiming

    Criticizes · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

27. I evaluated this project yesterday and found its claims un-credible. It's literally nothing like jev.

    Criticizes · Sep 2026 · in “I built non-autoregressive decision models with RL a year ago”

28. the paper contents are incomplete and describe something that does not sound like Jev at all

    Criticizes · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

29. Laya is a finetuned version of Google's BeRT model, which is almost 10 years old right now.

    Criticizes · Sep 2026 · in “Laya on Mac M4 CoreML Offline”

30. for the most part laya just fell apart and didn't return anything useful for a lot of stuff

    Criticizes · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

31. Laya came out after Jev and appears to be vibe coded to look more like Jev

    Criticizes · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

32. I evaluated this project yesterday and found its claims un-credible.

    Criticizes · Sep 2026 · in “I built non-autoregressive decision models with RL a year ago”

33. Zero-shot Laya tracks Jev on 2-4 labels and falls off a cliff on 77

    Criticizes · Sep 2026 · in “Does an open-weight decision model beat a hosted one? Jev vs. Laya”

34. Laya is nowhere close to Jev in the 3rd party benchmarks I've seen

    Criticizes · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

35. Laya is just ModernBert fine-tuned. It's still a language model

    Criticizes · Oct 2026 · in “Build your own decision model”

36. Setting up local decision / classification models is as easy as telling the agent you already use "set up laya on llama cpp locally"

    Recommends · Oct 2026 · in “We’re releasing Jev 1.14.0 along with experimental features for dreamers”

37. Laya is ModernBert-Large (421M) fine-tuned and has a similar to better performance than TypeSafe Jev

    Recommends · Oct 2026 · in “Build your own decision model”

38. having a model follow you instructions and be promotable and fast makes it massively easier to use

    Recommends · Sep 2026 · in “I built non-autoregressive decision models with RL a year ago”

39. this does seem undermined by the fact that Ollama can add support for decision models at any time

    Mixed · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”

40. The examples suffer the same problem of why I'd prefer to use a string column vs an enum.

    Mixed · Sep 2026 · in “Ollaya – Ollama for open-source, Jev-style decision models”