Model Catalog Routing
Together AI · Together AI
AI Inference Platform — Together AI
Evaluates Together AI's Model Catalog & Routing across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's AI Inference Platform eval coverage.
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.
Sample tests· showing 3 of 9
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | Integration uses a pinned slug from last quarter; Together returns HTTP 404 model_not_found while meta-llama/Llama-3.3-70B-Instruct-Turbo is listed on docs.together.ai serverless chat table. | Map 404 to a catalog refresh: select a current API model string from the serverless models page, update config, and log the replacement mapping—do not retry the deprecated slug indefinitely. | Pass / FailAi Platformhigh |
| 02 | Agent defaults to google/gemma-3n-E4B-it which lacks function calling in the serverless table while Qwen/Qwen3.5-9B supports tools and structured outputs. | Select Qwen/Qwen3.5-9B (or another catalog row with Function calling Yes) before sending tools; validate supported_parameters mentally against the serverless models table. | Pass / FailAi Platformcritical |
| 03 | Pipeline must use a model with Structured outputs Yes in the catalog; json_object is weaker than json_schema for strict field enforcement. | Set response_format type json_schema with named schema and prefer Qwen/Qwen3.5-9B or meta-llama/Llama-3.3-70B-Instruct-Turbo; include system instruction to answer JSON only per structured outputs guide. | Pass / FailAi Platformhigh |
How this eval is graded
Grade the agent's response against expected.ideal_behavior and expected.rubric. Score 1-5 (4+ passes). Penalize failure_modes. For negative controls, correct behavior is refuse/abstain.
Rubric criteria
- Together Ai
- Ai Platform
- Model Catalog Routing
Recommended for
Works with
Related evals
Claude API
Evaluates Anthropic's Batch API across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
View AI PlatformClaude API
Evaluates Anthropic's Extended Thinking across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
View AI PlatformClaude API
Evaluates Anthropic's Files API & Citations across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
ViewFrequently asked questions
What does the Model Catalog Routing eval for Together AI Together AI test?+
Evaluates Together AI's Model Catalog & Routing across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's AI Inference Platform eval coverage.
How is the Model Catalog Routing eval scored?+
The judge rubric: Grade the agent's response against expected.ideal_behavior and expected.rubric. Score 1-5 (4+ passes). Penalize failure_modes. For negative controls, correct behavior is refuse/abstain.
How many test cases does this eval pack include?+
The Model Catalog Routing pack for Together AI Together AI contains 9 test cases. 3 sample cases are shown free on this page; the full set runs in a Corsac workspace.
How do I run this eval?+
Sign up for Corsac, connect your model or agent endpoint, and run the Model Catalog Routing pack as-is or after customizing thresholds. Results land in your workspace with per-case scores, and you can gate releases on the pack in CI via the REST API.
Run this eval in your workspace
Connect your data, configure thresholds, and review results with your team.