the model

llama3.2:3b

Served by Ollama. Point a task at it and the audit prices the ceiling before a token is spent.

Vendored from the released binary · supernovae-st/nika@25f502adb7ec (v0.111.0) · digest-verified, re-derived at build, gated in CI

Point a task at it

tasks:
  ask:
    infer:
      model: "ollama/llama"
      prompt: "…"

The seats · 1

ollama/llama128k context · 8k out
Ollama · the provider room

The price

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

The bench

36.4s to answerollama/llama3.2:3b · 355 chars · 2026-08-03
One run, one machine: a witness, never a benchmark suite. The receipt carries the artifact pin, the evidence pack (chain-verified) and the spend under its announced ceiling · or run it yourself: the model-bench workflow in the public registry, swap the seats for the models you have.

{"id":"llama3.2:3b","served_by":[{"provider":"ollama","provider_name":"Ollama","alias":"llama","context_window_tokens":128000,"max_output_tokens":8192}],"slug":"llama3-2-3b","pricing":[],"energy":[]}