Eval Library
TA
For Together AIAI Platform

Inference Api Reliability

Together AI · Together AI

Together AI evals — Inference API Reliability (relift v3)

About Together AI

Together AI is an enterprise AI inference cloud providing fast, scalable access to leading open-source models via an OpenAI-compatible API. Teams use Together for production inference, fine-tuning, and dedicated GPU deployments without the complexity of self-managed infrastructure.

Employees

~100

Industry

AI Inference Platform

Headquarters

San Francisco, CA

Sample tests· showing 3 of 10

#InputExpected behaviorCheck
01

Client parses SSE chunks; stream must end with data: [DONE] per OpenAPI ChatCompletionStream.

Iterate chat.completion.chunk deltas until sentinel data [DONE]; concatenate delta.content; handle finish_reason on final chunk.

Pass / FailAi Platformmedium
02

Model must halt when emitting closing tag; stop is array of strings per schema.

Include stop list in request; verify finish_reason stop not length when tag emitted.

Pass / FailAi Platformhigh
03

503 indicates platform capacity not client over-limit; safe to retry with backoff unlike some mutations.

Retry POST with exponential backoff and jitter; read x-ratelimit-reset; do not disable safety_model on retry.

Pass / FailAi Platformcritical

Rubric criteria

  • Together Ai
  • Ai Platform
  • Inference Api Reliability

Recommended for

Together AITogether AI customers

Works with

Related evals

Run this eval in your workspace

Connect your data, configure thresholds, and review results with your team.