---
title: "Admin runtime benchmark catalog"
method: GET
path: "/v1/admin/runtime/benchmarks"
tags: ["System"]
---

# Admin runtime benchmark catalog

`GET /v1/admin/runtime/benchmarks`

## Request

- Base URL: `https://api.algenta.ai`
- URL: `https://api.algenta.ai/v1/admin/runtime/benchmarks`
- Auth: one of:
  - HTTP bearer
  - API key in header `X-API-Key`

## Response `200`

Successful Response

- RuntimeAdminBenchmarksResponse
  - `runtime_version` string, required
  - `llm_core_manifest` string, required
  - `module_manifest_version` string, required
  - `benchmark_discovery_lane` BenchmarkDiscoveryLane, required
    - `discovered_source_modules` integer, required
    - `discovered_public_functions` integer, required
    - `discovered_source_inventory` BenchmarkDiscoverySourceModuleEntry[]
      - `import_path` string, required
      - `public_function_count` integer, required
    - `shipping_manifest_modules` integer, required
    - `shipping_manifest_functions` integer, required
    - `non_shipping_rule` 'The broader Mojo inventory remains a benchmark and discovery lane only; modules are not advertised on the runtime surface unless listed in this manifest.', required
  - `benchmarking` BenchmarkFramework, required
    - `classes` BenchmarkClassEntry[]
      - `code` 'B1' | 'B2' | 'B3' | 'B4' | 'B5' | 'B6' | 'B7' | 'B8' | 'B9' | 'B10', required
      - `description` string, required
      - `evidence_paths` string[]
    - `required_metrics` RuntimeBenchmarkMetric[]
    - `baselines` RuntimeBenchmarkBaseline[]
  - `slo_budgets` SLOBudget[]
    - `name` 'runtime_manifest_load' | 'ttft' | 'end_to_end_turn' | 'agent_checkpoint_commit' | 'mcp_call_first_party' | 'decision_plan_creation' | 'replay', required
    - `applies_to` 'first-party runtime manifest route' | 'first-party native LLM serving path' | 'agent runtime checkpoint persistence' | 'first-party or internal MCP tools' | 'decision runtime plan generation' | 'runs up to 1000 events', required
    - `p95_objective_ms` number, required
    - `hard_ceiling_ms` number, nullable
    - `notes` string, nullable
  - `quality_gate_benchmark_classes` BenchmarkClassEntry[]
    - `code` 'B1' | 'B2' | 'B3' | 'B4' | 'B5' | 'B6' | 'B7' | 'B8' | 'B9' | 'B10', required
    - `description` string, required
    - `evidence_paths` string[]
  - `quality_gate_slo_budgets` SLOBudget[]
    - `name` 'runtime_manifest_load' | 'ttft' | 'end_to_end_turn' | 'agent_checkpoint_commit' | 'mcp_call_first_party' | 'decision_plan_creation' | 'replay', required
    - `applies_to` 'first-party runtime manifest route' | 'first-party native LLM serving path' | 'agent runtime checkpoint persistence' | 'first-party or internal MCP tools' | 'decision runtime plan generation' | 'runs up to 1000 events', required
    - `p95_objective_ms` number, required
    - `hard_ceiling_ms` number, nullable
    - `notes` string, nullable
  - `evaluation_science` EvaluationScience, required
    - `dimensions` RuntimeEvaluationDimension[]
    - `methods` RuntimeEvaluationMethod[]
    - `release_blockers` RuntimeReleaseBlocker[]
  - `evaluation_summary` RuntimeAdminEvaluationSummary, required
    - `dimension_count` integer, required
    - `method_count` integer, required
    - `release_blocker_count` integer, required
    - `benchmark_class_count` integer, required
    - `slo_budget_count` integer, required
    - `replay_gate_enabled` boolean, required
    - `tool_call_quality_gate_enabled` boolean, required
    - `rag_quality_gate_enabled` boolean, required
    - `decision_quality_gate_enabled` boolean, required
  - `compiled_artifacts` ArtifactReference[]
    - `kind` 'proof_bundle' | 'parity_benchmark' | 'compiled_runtime_benchmark' | 'compiled_mojo_binary', required
    - `path` string, required
    - `sha256` string, required
    - `size_bytes` integer, required
  - `shipping_runtime_modules` RuntimeModuleId[]
  - `shipping_runtime_function_counts` integer[]
  - `shipping_runtime_benchmark_speedups_x` number[]
  - `shipping_runtime_benchmark_artifacts` string[]
  - `shipping_runtime_compiled_artifacts` string[]
  - `shipping_runtime_compiled_engines` RuntimeCompiledEngine[]
  - `shipping_runtime_max_cold_ms` number[]
  - `shipping_runtime_max_warm_ms` number[]
  - `shipping_runtime_max_hot_ms` number[]
  - `module_benchmarks` RuntimeBenchmarkModuleEntry[]
    - `name` 'bpe_tokenizer' | 'text.tokenizer' | 'embeddings' | 'flash_attention' | 'sparse_attention' | 'transformer_attention' | 'transformer_blocks' | 'llm_sampling' | 'vector_similarity' | 'vector_search' | 'vector_kernels.ranker' | 'vector_kernels.table' | 'sparse_vector' | 'rlhf_dpo' | 'rerank_eval' | 'inference_cost_latency' | 'kv_cache' | 'paged_kv_cache' | 'continuous_batching' | 'inference_engine' | 'speculative_decoding' | 'generation_loop', required
    - `function_count` integer, required
    - `benchmark_artifact` string, required
    - `benchmark_speedup_x` number, required
    - `compiled_artifact` string, required
    - `compiled_engine` 'mojo' | 'python_fallback', required
    - `max_cold_ms` number, required
    - `max_warm_ms` number, required
    - `max_hot_ms` number, required
  - `manifest_digest` string, required
  - `signature` ManifestSignature, required
    - `algorithm` 'hmac-sha256', required
    - `key_id` string, required
    - `digest_hex` string, required
    - `signature_hex` string, required
    - `scope` 'control_plane_hmac_v1', required

---

[API](https://skmtc.dev/codna/apis/algenta-api.md) · [All operations](https://skmtc.dev/codna/apis/algenta-api/llms.txt) · [OpenAPI document](https://skmtc.dev/codna/apis/algenta-api/revisions/35d30f84a5df?raw)