Lafa scales flow language models to 8B, matching AR quality at 2.5x throughput
AdtRaghunathan · x · 2026-10-22
Researchers introduced Lafa, a flow language model scaled to 8B parameters, showing flow LMs can handle real generative tasks:
- Quality: matches autoregressive models at the 8B scale
- Speed: 2.5x AR throughput at 7B
A concrete data point that the non-AR text generation路线 can be competitive at scale.
Related event: Lafa scales flow language models to 8B with 2.5x throughput(3 posts)→
More from Models
- Artificial Analysis benchmarks 27 Search APIs across 14 providers; DeepSeek Search 3 points off the lead — ArtificialAnlys · 2026-10-22
- DeepSeek Search debuts at $0.75/1k searches, cheapest paid Search API, ranks 17th of 27 — ArtificialAnlys · 2026-10-22
- Jev 1.14 lands on Databricks, targeting fast decisions and complex tasks — matei_zaharia · 2026-10-22
- Reddit asks: which 50-150B MoE models are worth running locally besides Qwen Flash Next — FlyingFishMakeAWish · 2026-10-22
- Codex reset card timing trick: burn your quota before tomorrow's expected reset — xiaohu · 2026-10-22
- Verify TTS pronunciation with Kaldi phoneme decoding: a reproducible fix for polyphonic Chinese characters — dotey · 2026-10-22