# Blog

## What’s new at Fireworks

Blog

- **[Train Muse Glimmer for free: Our recipe for out-betting Jev](https://fireworks.ai/blog/train-muse-glimmer-for-free)**: One training pass, better predictions. See how we trained Muse Glimmer to outperform Jev on unseen Kalshi topics, then try the recipe with free SFT through November 30. · 20 OCT, 2026 · Announcement
- **[Introducing Ember-1](https://fireworks.ai/blog/ember-1)**: 23 SEP, 2026
- **[Introducing the Specialized Intelligence Index](https://fireworks.ai/blog/specialized-intelligence-index-ai-benchmarks)**: 22 SEP, 2026 · Announcement
- **[The frontier isn’t a model. It’s a router.](https://fireworks.ai/blog/llm-model-routing-deepswe)**: 21 SEP, 2026 · Perspective

1. **[Kimi K3.1 Pro is live on Fireworks](https://fireworks.ai/blog/kimi-k3-1-pro-fireworks)**: 4 NOV, 2026 · Model release
2. **[Fireworks unveils new products at Forge to help companies build specialized intelligence](https://fireworks.ai/blog/fireworks-unveils-new-products-at-forge-to-help-companies-build-specialized-intelligence)**: 3 NOV, 2026 · Announcement
3. **[Zeta meets the closed frontier at 94% lower cost and 6x speed with Fireworks-trained models](https://fireworks.ai/customers/zeta-fine-tuned-nemotron-models-case-study)**: 3 NOV, 2026 · Customer story
4. **[Paper Instruments trains a frontier-level model for knowledge work at ~98% lower inference cost](https://fireworks.ai/customers/paper-instruments-trains-a-frontier-level-model-for-knowledge-work-at-98-percent-lower-cost)**: 28 OCT, 2026 · Customer story
5. **[Heidi trains and deploys NVIDIA Nemotron 3 Ultra on Fireworks](https://fireworks.ai/customers/heidi-trains-and-deploys-nvidia-nemotron-3-ultra-on-fireworks)**: 22 OCT, 2026 · Customer story
6. **[Muse Spark Open on Fireworks: Meta’s open weights model, available on US serverless](https://fireworks.ai/blog/meta-muse-spark-open)**: 20 OCT, 2026 · Model release
7. **[Train Muse Glimmer for free: Our recipe for out-betting Jev](https://fireworks.ai/blog/train-muse-glimmer-for-free)**: 20 OCT, 2026 · Announcement
8. **[Kimi K3 on Fireworks: Fast RL at frontier scale](https://fireworks.ai/research/kimi-k3-on-fireworks-fast-rl-at-frontier-scale)**: 12 OCT, 2026 · Research
9. **[Reinforcement learning: Why alignment of numerics and MoE routing matter](https://fireworks.ai/blog/reinforcement-learning-why-alignment-of-numerics-and-MoE-routing-matter)**: 30 SEP, 2026
10. **[Inside Fireworks multi-region deployments: One deployment, global scale](https://fireworks.ai/blog/multi-region-deployment)**: 28 SEP, 2026
11. **[Every byte counts: ARCv3 and the case for cross-region RL](https://fireworks.ai/blog/arcv3-and-the-case-for-cross-region-rl)**: 23 SEP, 2026
12. **[Introducing Ember-1](https://fireworks.ai/blog/ember-1)**: 23 SEP, 2026
13. **[Introducing the Specialized Intelligence Index](https://fireworks.ai/blog/specialized-intelligence-index-ai-benchmarks)**: 22 SEP, 2026 · Announcement
14. **[The frontier isn’t a model. It’s a router.](https://fireworks.ai/blog/llm-model-routing-deepswe)**: 21 SEP, 2026 · Perspective
15. **[Phylo brings frontier AI to more scientists with open models on Fireworks](https://fireworks.ai/customers/phylo-biomni-lab-open-models-case-study)**: 17 SEP, 2026 · Customer story
16. **[DeepSeek-V4.1-Flash on Fireworks: Astra-level DeepSWE at 1/15th the cost](https://fireworks.ai/blog/deepseek-v4-1-flash-vs-astra-benchmarks)**: 14 SEP, 2026 · Model release
17. **[Gen-1 Slides: Opus 5-level decks at a fraction of the cost](https://fireworks.ai/customers/genspark-gen-1-slides-rl-post-training)**: 10 SEP, 2026 · Customer story
18. **[Making the leap to specialized intelligence](https://fireworks.ai/blog/specialized-intelligence-train-your-own-model)**: 9 SEP, 2026 · Learn
19. **[Train past the frontier: Training API now generally available](https://fireworks.ai/blog/training-api-general-availability)**: 31 AUG, 2026 · Announcement
20. **[DeepSeek V4 Pro: Tops SWE-Bench & cuts cost per task by 3x vs. Fable 5](https://fireworks.ai/blog/deepseek-v4-pro-vs-fable-5-benchmarks)**: 26 AUG, 2026 · Model release
21. **[DeepSeek V4 Pro is redefining security agent economics](https://fireworks.ai/blog/deepseek-v4-pro-cybergym-security-benchmark)**: 26 AUG, 2026 · Perspective

[More posts](https://fireworks.ai/blog/page/2)

## Build your own frontier.

Start running open models in minutes. Then orchestrate, train and climb toward intelligence no one else can rent.

[Get started](https://app.fireworks.ai/signup) · [Talk to experts](https://fireworks.ai/contact)