Batch API
Send an OpenAI Batch JSONL file and get completions from a pinned Hugging Face model, with the OpenAI SDK, plain HTTP or an AI agent.
Explore 45 official skills across compute, observability, resources, jobs, generation, evaluation and training through the customer CLI, MCP or direct HTTP API. Public reads need no key. Execution remains gated.
Install CLI 0.12.0 · Connect MCP · HTTP API
Send an OpenAI Batch JSONL file and get completions from a pinned Hugging Face model, with the OpenAI SDK, plain HTTP or an AI agent.
Score a pinned model on held-out sets such as AIME, GPQA and MMLU-Pro, or on your own questions with reference answers.
Check the configured client, API origin and service health without treating configured keys as validated access.
Check my Reliquary connection and explain what I can use next.
reliqua doctor --jsonCompare GPU provider observations with whole-package prices, source freshness and booking limits.
Compare H200 offers, preserve their sources and freshness, and explain what I can actually book.
reliqua compute list --gpu H200 --limit 20 --jsonIdentify published checkpoints and immutable revisions while keeping public observations separate from private training readiness.
What is the latest published checkpoint, and does it show that my training can run today?
reliqua inventory --jsonRead current product catalogs and readiness gates before choosing a supported workflow.
Which Reliquary products are currently available, gated or only observable?
reliqua capabilities --jsonInspect collection health and the published monitoring snapshot to explain missing or stale observations.
Is Reliquary monitoring fresh, and which observations are missing?
reliqua monitoring health --jsonCompare published network-wide observations over a supported 24-hour, 7-day or 30-day range.
Compare the published network history over the last seven days and show gaps.
reliquary_connection → reliquary_monitoring_historyRead one supported public analytics view with bounded filters and source-specific metric meanings.
Explain the published training analytics and what they actually measure.
reliquary_connection → reliquary_monitoring_analyticsInspect qualified generation models, sources, environments, accepted model architectures and current ordering limits before requesting a dataset.
Which dataset generation sources and models can I use?
reliqua generation catalog --jsonRequest a generation estimate for exact catalog inputs or a pinned public model without ordering work.
Quote a small dataset generation request and show the maximum cost before I decide.
reliquary_connection → reliquary_generation_catalog → reliquary_generation_quote → reliquary_generation_pinned_quoteSubmit one exact generation quote after approval, preserving its cost ceiling and request identity.
Place the dataset order from this quote after I approve its exact cost and limits.
reliquary_connection → reliquary_generation_catalog → reliquary_generation_orderRead owned generation orders and explain progress, delivery and refund state without polling.
What happened to my dataset order, and has it delivered?
reliquary_connection → reliquary_generation_orders → reliquary_generation_order_getRetrieve a bounded range of an owned delivered dataset using identifiers and integrity metadata returned by the order.
Show the delivered dataset files for my order and retrieve a reviewed sample.
reliquary_connection → reliquary_generation_order_get → reliquary_generation_downloadRequest cancellation of one owned generation order after showing current progress and refund uncertainty.
Cancel my dataset order after reviewing what can still stop and be refunded.
reliquary_connection → reliquary_generation_order_get → reliquary_generation_cancelInspect qualified evaluation sets, supported execution modes and current ordering limits.
Which held-out evaluation sets and modes are available?
reliqua evaluations catalog --jsonRequest a held-out evaluation quote for an immutable public model and exact frozen-set limits.
Estimate the cost of evaluating this exact model revision on supported held-out sets.
reliquary_connection → reliquary_evaluation_catalog → reliquary_evaluation_quoteSubmit one reviewed evaluation quote after approval and keep execution, delivery and model quality separate.
Run this evaluation only after I approve the quote and its limits.
reliquary_connection → reliquary_evaluation_catalog → reliquary_evaluation_orderInspect one owned evaluation and its bounded reports while checking frozen sets, grading completeness and comparison limits.
Review my evaluation results and tell me what they establish about this model.
reliquary_connection → reliquary_evaluations → reliquary_evaluation_get → reliquary_evaluation_downloadRequest cancellation of one owned evaluation after reviewing current state and refundable-credit limits.
Cancel this evaluation and explain any work or cost that remains.
reliquary_connection → reliquary_evaluation_get → reliquary_evaluation_cancelRead scoped training capabilities, billing mode, upload permission and qualified runtime profiles for the current account.
Can my account train today, and which exact runtime profiles are qualified?
reliquary_connection → reliquary_training_capabilitiesInspect owned training files and upload compatible bounded JSONL only after confirming data rights and current upload permission.
Inspect my owned JSONL dataset or prepare a compatible upload without starting training.
reliquary_connection → reliquary_training_capabilities → reliquary_training_files → reliquary_training_file_download → reliquary_training_uploadImport one verified owned generation JSONL shard into the training library without converting formats or starting a run.
Import this delivered dataset shard into my training library after checking ownership and rights.
reliquary_connection → reliquary_training_capabilities → reliquary_generation_order_get → reliquary_training_import_deliveryCreate a schemaVersion 2 training configuration from an exact current runtime profile and owned compatible dataset.
Save a bounded training draft using a qualified profile and my existing dataset.
reliquary_connection → reliquary_training_capabilities → reliquary_training_files → reliquary_training_draft_create → reliquary_resource_getQuote the current saved training draft revision and owned dataset in service-credit billing mode.
Quote this saved training draft and show its allocation and cost limits.
reliquary_connection → reliquary_training_capabilities → reliquary_resource_get → reliquary_training_quoteLaunch an exact quoted training draft only after approval of its runtime, dataset, revision and cost ceiling.
Start this exact training run after I approve the current quote and configuration.
reliquary_connection → reliquary_training_capabilities → reliquary_resource_get → reliquary_training_quote → reliquary_training_startRead owned training runs, pinned configuration, events and billing on a requested progress update.
Show my training run status and explain its latest events and billing.
reliquary_connection → reliquary_training_tasks → reliquary_training_task_getInspect and read bounded ranges of owned training artifacts using IDs and provenance returned by task details.
List my completed training artifacts and retrieve the requested report.
reliquary_connection → reliquary_training_task_get → reliquary_training_artifact_downloadRequest cancellation of one owned training task after explaining current execution and billing state.
Cancel this training run and report whether work and billing have actually stopped.
reliquary_connection → reliquary_training_task_get → reliquary_training_cancelInspect, archive or restore an owned training-library file while preserving active-run and quote dependencies.
Archive this unused dataset, or restore it if I need it again.
reliquary_connection → reliquary_training_files → reliquary_training_file_archive → reliquary_training_file_restoreRead the relevant scoped service-credit balance and ledger without conflating customer credits with network payouts.
Show the credit balance and ledger for my training or dataset order.
reliquary_connection → reliquary_generation_balance → reliquary_evaluation_balance → reliquary_training_balanceInspect and edit owned saved resource configurations with exact revision control and no implied execution.
Find my saved draft and update its reviewed configuration without launching it.
reliquary_connection → reliquary_resources_list → reliquary_resource_get → reliquary_resource_create → reliquary_resource_update → reliquary_resource_archiveInspect owned CPU-job permission, fresh qualified workers and bounded backlog without treating permission as available capacity.
Can my account run a small CPU dataset validation job now?
reliqua jobs capabilities --jsonUpload one bounded owned JSONL input and submit an explicitly approved CPU validation specification without paid provisioning.
Validate this small dataset on CPU after I approve its upload and bounded job.
reliquary_connection → reliquary_jobs_capabilities → reliquary_job_input_upload → reliquary_job_submitRead owned CPU jobs and bounded events while preserving real work-unit state, attempts and uncertainty.
Show the current state and latest events for my CPU validation job.
reliqua jobs list --limit 20 --jsonRead a frozen owned CPU report and verify its size and digest against job metadata before interpreting row/error counts.
Retrieve my CPU validation report and explain which checks passed.
reliqua jobs report --id JOB_ID --artifact ARTIFACT_ID --jsonControl one owned CPU validation job at its exact reviewed revision after approval, preserving request identity.
Pause or cancel this CPU validation job after reviewing its current revision.
reliquary_connection → reliquary_jobs_capabilities → reliquary_job_get → reliquary_job_controlFind the current or historical public run to inspect, using published identifiers and availability.
Which public training runs can I inspect, and which one is current?
reliqua runs list --jsonRead stored public trainer and controller observations, keeping checkpoint progress, collection freshness and missing data distinct.
Is the current RL run progressing, and which observations are fresh?
reliqua runs inspect --jsonCompare published held-out evaluation observations from a public run, preserving the set, checkpoint, sample count and comparison limits.
How did the published benchmark results change from the baseline?
reliqua runs inspect --jsonInspect actual raw response-token volume by environment from complete trainer observations, separately from prompt counts and PPO weights.
Show the actual response-token mix by environment for the latest complete observation.
reliqua runs inspect --jsonCompare public GPU model participation and qualified quality windows while preserving declarations, unavailable metrics and source clocks.
Which declared GPU models are participating, and is their quality comparison qualified?
reliqua runs inspect --jsonReview a public MiMo-compatible checkpoint, approve its bounded preparation and explicitly launch one source-pinned diagnostic.
Prepare my compatible checkpoint for the assisted Affine pilot and show each approval boundary.
reliquary_connection → reliquary_affine_pilot_status → reliquary_affine_checkpoint_review → reliquary_affine_checkpoint_prepare → reliquary_affine_preparation_get → reliquary_affine_evaluation_prepare → reliquary_affine_request_upload → reliquary_resource_create → reliquary_training_quote → reliquary_training_start → reliquary_training_task_get → reliquary_training_artifact_downloadReview a v1/v2 batch, prepare a portable plan, or prepare an approved owner-scoped hosted batch request.
Prepare batch generation and export, official tokenizer validation, one GRPO step and parent/candidate comparison.
reliqua affine status --jsonUpload an OpenAI Batch JSONL file, quote it on a pinned model and order it after explicit approval.
Run my batch file requests.jsonl on this model and show me the quote before ordering.
reliquary_connection → reliquary_generation_catalog → reliquary_batch_upload → reliquary_batch_file_get → reliquary_batch_quote → reliquary_batch_orderWrite every delivered file of an owned order to a local directory with sha256 verification, without loading data into the conversation.
Download the results of my batch order into ./results.
reliquary_connection → reliquary_batch_files → reliquary_batch_downloadEvaluate a model on your own questions with reference answers: a Batch JSONL file graded by boxed_math_v1 or exact_match_v1, quoted on a pinned model and ordered after explicit approval.
Evaluate Qwen/Qwen3-8B on the questions and reference answers in eval.jsonl with boxed_math_v1, and show me the quote before ordering.
reliquary_connection → reliquary_generation_catalog → reliquary_batch_upload → reliquary_batch_file_get → reliquary_batch_quote → reliquary_batch_order → reliquary_generation_order_get → reliquary_batch_files → reliquary_batch_downloadDownloads, checksums and compatibility
Current catalogs control availability. A listed offer or published checkpoint does not qualify execution.