saricles/qwen3-coder-next
qwen3-coder-next (eugr)
Throughput
63.3t/s@ 256k
Engine
vLLM
Parameters
3B active / 80B total
Benchmarked
2026-06-27
Context ladder
Throughput at each benched context window (single measurement).
| Context | KV | Throughput |
|---|---|---|
| @ 256k peak golden | — | 63.3t/s |
Golden profile
saricles-qwen3-coder-next-eugr
Capabilities
labeugrgoldenvllm
Why we run it
Golden fleet target — auto-scaffolded from recipe saricles-qwen3-coder-next-eugr.
Bench notes
golden ?/? @ 56.8 tok/s — fill~50000 — bench-agent-v2 — tool_ok=False
Links
- saricles/qwen3-coder-next — gated or private repo on HuggingFace, no public link available.
- Golden recipe definition ↗