the model

Qwen/Qwen3-8B

Served by Native Inference · vLLM. Point a task at it and the audit prices the ceiling before a token is spent.

Vendored from the released binary · supernovae-st/nika@5d61b7e13f1f (v0.107.0) · digest-verified, re-derived at build, gated in CI

Point a task at it

tasks:
  ask:
    infer:
      model: "native/qwen"
      prompt: "…"

The seats · 2

native/qwen128k context · 8k out
Native Inference
vllm/qwen128k context · 8k out
vLLM · the provider room

The price

No exact-match row at this pin · the engine resolves pattern rules at run time and nika audit prints the real ceiling. A local seat is unpriced, never free.

{"id":"Qwen/Qwen3-8B","served_by":[{"provider":"native","provider_name":"Native Inference","alias":"qwen","context_window_tokens":128000,"max_output_tokens":8192},{"provider":"vllm","provider_name":"vLLM","alias":"qwen","context_window_tokens":128000,"max_output_tokens":8192}],"slug":"qwen-qwen3-8b","pricing":[],"energy":[]}