Fireworks offers free Muse Glimmer 30B SFT, beats Jev by 11.2 points on prediction accuracy
sophiamyang · x · 2026-10-22
Fireworks launched its Free Weights Club with free supervised fine-tuning of Muse Glimmer 30B through November 30. Fine-tuned on historical Kalshi markets, the model hit 58.3% prediction accuracy on unseen topics, 11.2 points above Jev's 47.1%, and a backtest turned a hypothetical $1,000 into $1,071 after fees while Jev and the base model lost money. They also released a Kalshi cookbook, and their System One serving stack ran 2-3× faster than standard BF16 chat completions on a single B200. Managed Training and Serverless Training API are available for custom SFT jobs.
Related event: Fireworks Offers Free Muse Glimmer 30B Fine-tuning for Kalshi Prediction(3 posts)→
More from Infra
- Google Cloud launches Agent Compute Engine, a new VM type built for AI agents — testingcatalog · 2026-10-23
- LM Studio ships Read Aloud, powered by tiny local TTS model Phonon from Gradium — mattturck · 2026-10-23
- Lam Research system revenue up 35% QoQ as AI pushes 2027 WFE past $200B — BenBajarin · 2026-10-23
- Hugging Face Hub launches OpenUSD support with 2,500+ NVIDIA SimReady assets — mishig25 · 2026-10-23
- One lying fsync() on a single node can make a Raft cluster lose data — tianyin_xu · 2026-10-23
- Hacked DFlash drafter runs on stock llama.cpp, +61% code speedup for Ling-3.0-tiny — Barni275 · 2026-10-23