Code & Cocktails in SF with Fireworks & MiniMax

Join Fireworks and MiniMax for an evening of great drinks, builder networking, and a fireside chat about what it takes to build open models for agents that run long, multi-step work.

MiniMax M3.1 Flash is live on Fireworks, and we’re celebrating with the team behind it!

About M3.1 Flash
Released as open weights, and available on Fireworks from day 0:

→ Cowork automation: 62.8 on AutomationBench-AA
→ Coding: 73.7% on DeepSWE v1.1
→ Visual reasoning: 91 on BabyVision
→ $0.15 input $0.60 output $0.015 cached per 1M tokens, half the price of M3

The fireside chat
Sinan Ozdemir, Head of AI Developer Education at Fireworks sits down with Ryan Lee, Head of Devrel at MiniMax to talk about:

• How M3.1 Flash was built for long-running agent workflows

• What makes a model work for coding, tool use and visual reasoning at the same time

• Why cheaper cached tokens change the economics of agent loops

• What’s next for open models

Agenda
6:00 PM · Doors open, drinks and bites
6:45 PM · Fireside chat + audience Q&A
7:00 PM · Networking
8:30 PM · Close

Location will be shared after registration is accepted.

About us:
Fireworks is the platform for specialized intelligence, enabling companies such as Doximity, Revolut, and Shopify to train and serve models tailored to their own data, workflows, and use cases. Founded by the team behind PyTorch, and backed by AMD, Atreides, Benchmark Capital, Index Ventures, Lightspeed, NVIDIA, Sequoia Capital, and TCV, Fireworks provides hundreds of state-of-the-art open models in text, image, embedding, and multimodal formats globally.