models

choose a language model

Start with brok-2.3. Use brok-2.3-fast when interactive latency matters more than maximum depth.

• brok-2.3flagship · 1,000,000-token context · reasoning, tools, streaming, vision, and video understanding
• brok-2.3-fastlow latency · 204,800-token context · interactive apps, agents, tools, and streaming
• brok-2.2stable previous generation · 204,800-token context
• brok-2.1legacy generation for existing workloads

generation models

• brok-image-1image generation
• brok-video-3latest video generation
• brok-video-2.3alternate video generation
• brok-speech-2text to speech with streaming
• brok-music-3music generation

stable names

brok-latest and brok-reasoning select the flagship model. brok-fast selects the low-latency model. brok-coding selects the flagship for coding work.

models = client.models.list()
for model in models.data:
    print(model.id)

Authenticated GET /v1/models is the source of truth for the current catalog.