All systems operational
Live status and speed of the Umans Code gateway and its models, refreshed every 30 seconds. For each model we show its median time to first token and output speed.
API endpoint
Models in production
TTFT = time to first token · tok/s = output tokens per second · p50 = median over the last 5 minutes. On each gauge the midpoint is that model's target; the marker sits further right when it's beating target (faster TTFT, higher throughput).
Umans Kimi K2.7 Code Recommendedsuccessor to Kimi K2.6also served as umans-coderOperational1.76sTTFT · p5035.1tok/sthroughput · p5098.12%uptime · 24h
Kimi K2.7-Code via Umans Code - Moonshot's strongest coding model and the successor to Kimi K2.6. Built for complex, tool-heavy agentic coding; it reasons more efficiently than K2.6, so agent sessions run faster at the same depth.
Umans Flash Fastestalso served as umans-qwen3.6-35b-a3bOperational1.03sTTFT · p50169.8tok/sthroughput · p5099.93%uptime · 24h
Our fastest model for coding-agent workloads. Optimized for low-latency iteration with tools and coding assistants like Claude Code.
Umans GLM 5.2 Operational4.05sTTFT · p5065.8tok/sthroughput · p5099.72%uptime · 24h
GLM 5.2 is our best model for coding right now, with a 400K context window for large codebases. Vision is available on the Anthropic Messages API (`/v1/messages`) only, through a server-side handoff (GLM 5.2 generates the text, Kimi preprocesses the image); that handoff will be retired soon in favour of more efficient client-side image handling.
In the playground
Umans GLM 5.2 NVFP4 ExperimentalIn testing…TTFT · p50 · gathering data…throughput · p50 · gathering dataJun 29in testing
Experimental NVFP4-quantized build of GLM 5.2, open for a short test window starting June 29, 2026 - temporary, not a permanent model. It isn't reachable before then: selecting it earlier returns errors (HTTP 500), so come back on the 29th. During the window - the published NVFP4 results look very flattering, but we're skeptical of its real quality and performance: this checkpoint was NOT QAT post-trained for NVFP4 (the QAT-on-NVFP4 models are where we've seen the best quality), so treat the benchmarks with caution. Play with it, push it, and see how far it gets you - for production work we recommend the fp8 `umans-glm-5.2`. It's offered at low capacity and low availability, so expect it to be flaky and to go down under load: crash it, give it a moment, and try again.