←Story detail

Trending story

Fireworks AI opens free SFT training for Muse Glimmer 30B

2 reports2 story updatesUpdated Oct 20 08:00

Understand this story

AI overview

Fireworks AI announced that Muse Glimmer 30B is available for free SFT training until November 30, and it has published a complete recipe for prediction experiments using historical Kalshi market data. According to the results it published, the fine-tuned Glimmer achieved 58.3% prediction accuracy on unseen topics, 11.2 percentage points higher than Jev's 47.1%; in a historical backtest, a hypothetical $1,000 grew to $1,071 after simulated fees, while base Glimmer and Jev both lost money. Training used LoRA adapters and a K-way cross-entropy loss, ran through the training data only once, and separately used a calibration set to fit a temperature parameter for each question category. The same release also mentioned that in another synthetic benchmark, Fireworks' System One serving stack had a median server processing speed 2-3 times faster than standard BF16 chat completions. The free training window currently runs until November 30.

AI-generated from public reports

Latest development · Report

Fireworks Offers Free Training of Muse Glimmer 30B: Using SFT to Predict Kalshi Market Outcomes

Developments

1 confirmed developments

  1. Fireworks AI opens free SFT training for Muse Glimmer 30B

    Source:Fireworks AI

Reports and discussion timeline

Browse reports and discussion by time. Their evidence levels remain distinct.

Oct 20

  1. Fireworks AIFirst-party

    Fireworks Offers Free Training of Muse Glimmer 30B: Using SFT to Predict Kalshi Market Outcomes

    Fireworks AI announced that Muse Glimmer 30B can be trained with SFT for free until November 30, and released a complete recipe for a prediction experiment using historical Kalshi market data. The fine-tuned Glimmer achieved 58.3% accuracy on unseen topics, 11.2 percentage points higher than Jev's 47.1%; in a historical backtest, a hypothetical $1,000 grew to $1,071 after accounting for simulated fees, while both base Glimmer and Jev lost money. Training used LoRA adapters and a K-way cross-entropy loss, ran through the training data only once, and separately used a calibration set to fit a temperature parameter for each question category.

  2. Fireworks AIFirst-party

    Fireworks Free Weights Club: Training Muse Glimmer to Beat Jev on Predictions

    Fireworks AI launches the Free Weights Club, offering free SFT training for Muse Glimmer 30B until November 30. The fine-tuned tuned Glimmer achieves 58.3% prediction accuracy on unseen topics, 11.2 percentage points higher than Jev's 47.1%; in a historical backtest, after deducting simulated trading fees, a hypothetical $1,000 grew to $1,071, while both Jev and base Glimmer posted losses. In a separate synthetic benchmark, Fireworks' System One serving stack delivered median server throughput 2-3x faster than standard BF16 chat completions.

Related stories

  1. Related storiesEnglish translation is being prepared.
  2. Related storiesSGLang natively supports and powers the Beam production API
  3. Related storiesDeepSeek releases DeepSeek-V4.1-Flash-1017
  4. Related storiesMeta releases open-weight model Muse Spark Open