Arcee AI: Virtuoso Large

POST/v1/chat/completions

Virtuoso‑Large is Arcee's top‑tier general‑purpose LLM at 72 B parameters, tuned to tackle cross‑domain reasoning, creative writing and enterprise QA. Unlike many 70 B peers, it retains the 128 k context inherited from Qwen 2.5, letting it ingest books, codebases or financial filings wholesale. Training blended DeepSeek R1 distillation, multi‑epoch supervised fine‑tuning and a final DPO/RLHF alignment stage, yielding strong performance on BIG‑Bench‑Hard, GSM‑8K and long‑context Needle‑In‑Haystack tests. Enterprises use Virtuoso‑Large as the "fallback" brain in Conductor pipelines when other SLMs flag low confidence. Despite its size, aggressive KV‑cache optimizations keep first‑token latency in the low‑second range on 8× H100 nodes, making it a practical production‑grade powerhouse.

Request body

Parameters supported by this model. Values, defaults, and limits are read from the model registry.

stringmodelrequired

Model identifier. Set to arcee-ai/virtuoso-large.

Default: arcee-ai/virtuoso-large

array<object>messagesrequired

Conversation messages in system, user, or assistant order.

Optional<integer>max_tokens

Maximum number of tokens the model may generate in the response.

Range: −∞ to 64000

Optional<number>temperature

Sampling temperature. Lower values are more deterministic; higher values are more creative.

Range: 0 to 2

Default: 1

Optional<number>top_p

Nucleus sampling threshold. Use this or temperature, but usually not both.

Range: 0 to 1

Optional<boolean>stream

When true, returns incremental Server-Sent Events instead of one completed response.

Default: false

Optional<array>tools

Tool definitions that the model may call during the response.

Optional<string>tool_choice

Controls whether the model may call a tool and, when supported, which tool it must call.

Optional<array<string>>stop

Sequences that stop generation when the model produces one of them.

Optional<number>presence_penalty

Penalizes tokens that already appeared, encouraging the model to introduce new topics.

Range: -2 to 2

Optional<number>frequency_penalty

Penalizes repeated tokens, reducing repetition in the generated response.

Range: -2 to 2

Optional<integer>top_k

Request parameter supported by this model.

Range: 0 to ∞

Response Schema

Fields returned by this model API response.

array<object>choicesrequired

Generated completion choices.

stringidrequired

Unique chat completion identifier.

stringmodelrequired

Model that generated the response.

Optional<object>usage

Token usage when available.

Model capabilities

array<string>capability_tagsrequired

Capabilities declared by the model registry.

Default: chat, function_calling

integercontext_lengthrequired

Maximum context window accepted by this model.

Default: 131072 tokens

stringexecution_moderequired

Execution mode declared by the model registry.

Default: sync

curl -X POST https://api.sandbase.ai/v1/chat/completions \
  -H "Authorization: Bearer $SANDBASE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "arcee-ai/virtuoso-large",
  "messages": [
    {
      "role": "user",
      "content": "Describe this product in one sentence."
    }
  ]
}'

Response200 OK

{
  "id": "chatcmpl_abc123",
  "model": "arcee-ai/virtuoso-large",
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "A concise product description."
      }
    }
  ]
}