Model coverage
One agreement covers the catalog.
Two lines with very different economics: an exclusive flagship line, and the open-weight catalog your teams already know — same contract, same controls.
Closed commercial Qwen, only from the publishing platform.
These publish no weights, so they are served only by the platform that owns them — the same platform this service runs on. Hosted-inference providers structurally cannot offer them.
qwen3.7-maxCurrent flagship of the closed commercial Qwen line. Weights are not published; access is API-only through Alibaba Cloud Model Studio.
qwen3.7-plusThroughput-oriented sibling of qwen3.7-max in the closed commercial line.
qwen3.8-max-previewPreview of the next closed commercial flagship; API-only from day one.
GLM, MiniMax, DeepSeek and Kimi access.
Open-weight is the precise term: the weights are public, the training data and code are not. Operationally what matters is that they run behind the same agreement, capacity controls and reporting as the flagship line.
qwen3.6-27bThe most recent open-weight general Qwen model.
deepseek-v4DeepSeek’s current open-weight flagship.
deepseek-v4-flashLatency/cost-optimized DeepSeek V4 variant with heavy production adoption.
glm-5.1Z.ai’s current open-weight GLM release.
kimi-k2.7-codeMoonshot’s coding-focused K2-series model.
minimax-m3MiniMax’s open-weight M3, a high-volume production workhorse.
Model availability varies by provider region.
Upstream regions run independent catalogs, keys and capacity pools. Which models are available where — and at what throughput — belongs in the pilot scoping conversation, not in a static list we would let go stale. Current status lives on the availability ledger.
Tell us which models matter.
We scope pilots around your model list, region needs and monthly volume.
hello@steadygateway.com99.9% availability commitment with tiered service credits · Reply within one business day · NDA available on request