Skip to content

Maxwell V1.0 Flash

Coming soon

Coming soon. The Maxwell team published Apache-2.0 weights on 2026-11-13; no inference API with published token prices serves them yet. We will list prices and route requests once a host does.

maxwell-model/maxwell-v1-flash

Text · 262K context · Nov 2026 · 0 upstreams

Maxwell V1.0 Flash is the Maxwell team's open-weight 35B/3B-active sparse MoE with a 256K-token context. It returns text for agentic work, grounded research and software engineering. Listed as upcoming.

Coming soon

Announced, not routable yet. Follow the changelog for the launch.

context
262K
announced price
TBA
Aliases (2)
  • maxwell/maxwell-v1-flash
  • maxwell-model/Maxwell-V1-Flash
Text generation
Context length262,144 tokens
Max output209,715 tokens
ArchitectureTransformer
Categorytext
ReleasedNov 13, 2026
Modalities
Capabilities

Gallery

Images from the vendor
  • Maxwell V1.0 Flash performance overview (Maxwell team-reported)© Maxwell team