you are viewing a single comment's thread
view the rest of the comments
[–] 12 points 5 days ago* (last edited 5 days ago) (13 children)

But is it actually better than Ornith or Tiel 35b-a3b?

I'm looking at around 30-50k context until I start running out of room.

I can do 80k+ just fine with 35b-a3b. I don't think 50k would be enough for agentic coding.

  • source
  • hideshow 13 child comments
  • [–] 7 points 5 days ago (7 children)

    I run ornith 1.5 35b a3b on my 8gb Rx 7600, get 30 t/s, and have context at 130k. It slows down to around 15 t/s near the end on the context window.

    Nothing I have tested comes close to this model.

    I use it in pi with search, coding sandbox and limited filesystem access.

  • source
  • parent
  • hideshow 7 child comments
  • [–] 1 point 5 days ago (4 children)

    Benchmarks like SWE-Bench Pro are pretty close to Qwen 3.8 27B. I suppose this model is about what a Qwen 3.8 35b-a3b model would be. Alibaba appears to have moved onto Qwen 4 and haven’t announced any 35b-a3b models, so I’m glad this exists.

  • source
  • parent
  • hideshow 4 child comments
  • [–] 4 points 5 days ago* (last edited 5 days ago) (3 children)

    I don't really understand your comment? Qwen3.8 27b is qwen4 architecture as a pre-release. apparently i was wrong here, it's still on the 3.5 architecture.

    A 27b dense model is really difficult to compare to a 35b a3b Moe model. I tried to run qwen3.8 27b on my 8gb card and ok it runs technically, but in no way is it viable or does it come close to ornith 1.5 35b a3b, mostly because to run qwen3.8 27b, I have to choke it to almost useless levels by running an aggressive quant like a 1, 2, or 3 bit quant, and even then it doesn't run as fast.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 4 points 5 days ago (2 children)

    FYI: Qwen3.8 27B is still a Qwen3.5 architecture model (Qwen3_5ForConditionalGeneration in config.json); Qwen3.8-Flash-Next -- a different model, despite the similar name -- is the Qwen4 preview (Qwen4ExpForConditionalGeneration).

    I believe what they're saying is that since we didn't get a MoE around the 35B-A3B size for the Qwen3.8 generation of releases that this model might be able to fill that niche -- i.e. a better model than Qwen3.6 35B-A3B while still being less compute intensive to run than Qwen3.8 27B.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 5 days ago

    Yeah, exactly, and sorry to everyone if I was unclear. I’m not holding out hope for a Qwen 4 35b-a3b either, so it’s good to see a model that comes close to 3.8 27b in benchmarks and I hope they stay competitive with Qwen 4 27b as well.

  • source
  • parent
  • [–] 6 points 5 days ago (2 children)

    And you too have 8GB of VRAM? Could you please tell me where can I find such configuration to try?

  • source
  • parent
  • hideshow 2 child comments
  • [–] 11 points 5 days ago* (last edited 5 days ago) (1 child)

    Yeah RTX 2080 8GB and 32GB of system RAM, I wrote a guide

    https://lemmus.org/post/24235317

    I'll have to do an updated guide once Llama.cpp v0.7.0 comes out because there's a new option now for MoE expert caching. It doesn't work that well yet but I think they'll continue to improve it. https://lemmus.org/post/25955310

  • source
  • parent
  • hideshow 1 child comment
  • [–] 5 points 5 days ago (1 child)

    Ormit? Tiel? Can you elaborate? Are they based on qwen 35b A3B?

  • source
  • parent
  • hideshow 1 child comment
  • [–] 4 points 5 days ago*

    Yes they're based on Qwen. Ornith 1.5 is a fine-tune of Qwen.

    Tiel Coder is based on Ornith 1.5, it's technically just a quant and an improved template, but it seems like an improvement for people who are going to use Q4 anyways.

  • source
  • parent