# Celeris-1 Decision > Celeris-1 Decision is a hosted multimodal diffusion model from Celeris, a Marqo Inc research lab. Its HTTP API answers typed questions about text, structured data and images with probabilities and optional explanations. - Canonical: https://www.anchorterminal.com/tools/celeris-1-decision - Markdown: https://www.anchorterminal.com/tools/celeris-1-decision.md (~5,400 tokens) - Slim: https://www.anchorterminal.com/tools/celeris-1-decision.min.md (~1,580 tokens, same facts, less prose, for token-sensitive contexts) - JSON: https://www.anchorterminal.com/tools/celeris-1-decision.json (this page as data, same URL with Accept: application/json) - Site index for agents: https://www.anchorterminal.com/llms.txt (full text: https://www.anchorterminal.com/llms-full.txt) - API: https://www.anchorterminal.com/api/v1/index.json - Updated: 2026-10-09 ## Overview **Grade D · 49.3/100 · rank #708 of 842 · #10 in Decision models · not agent-ready · confidence medium** ## Assessment Typed probabilities, image inputs and optional explanations suit routing and classification inside an agent. Both System One and OpenAI Decisions request formats are documented. The service remains early access under its terms, activation can queue, and prepaid credit expires after 30 days. The published accuracy and latency figures are Celeris measurements, not Anchor Terminal tests. ## Facts | Field | Value | | --- | --- | | Vendor | Celeris (Marqo Inc) (https://celeris.ai) | | Kind | Model API | | Category | Decision models (https://www.anchorterminal.com/categories/decision-models) | | Transport | HTTP | | Endpoint | `https://inference.celeris.ai/celeris-1-decision/v1/systemone` | | Auth | API key · Bearer ck_ API key from an activated workspace. Signup, a payment method and prepaid credit are required. Keys can be created, rotated and revoked in the console; changes can take one minute. Keys share workspace credit, with no per-key limits. Activation is usually immediate but can queue when capacity is full. | | Pricing | Pay per use (Pay per use) · $0.04 per million input tokens, including cached input, with output tokens free. Images and request overhead count as input. Card-funded purchases start at $5 and credit expires after 30 days. No card-free tier was found. Requests still processing after disconnection may be charged. Executed order forms can override self-service terms. | | x402 | No · No x402, MPP or L402 interface found in the reviewed pricing and API documentation on 9 October 2026. HTTP 402 means exhausted prepaid credit. | | Licence | Proprietary hosted model under Celeris terms of service | | Docs | https://docs.celeris.ai/decisions | | llms.txt | https://docs.celeris.ai/llms.txt | | Last release | 2026-10-08 | | Surface graded | The hosted celeris-1-decision API. System One at /v1/systemone and OpenAI Decisions format at /v1/decisions share the model and rate limits. | | Question types | System One uses noul, choice and score. The Decisions format uses predicate, choice and score. Answers include probabilities; explanations are optional. | | Request limits | 512 questions, reduced to 64 with explanations; 512 choices; 64 score levels; eight images of at most 25 megapixels each; 8 MiB request body. | | SDK compatibility | Celeris documents the Jev Python SDK for System One and OpenAI Python SDK 3.26.0 or later for Decisions. These are compatible clients from other vendors, not Celeris-owned packages. | | Context and streaming | No exact decision-model token window found. The 131,072-token window on the models page belongs to the separate chat models. Decision responses do not stream. | | Rate limits | Shared per workspace and model, with 429 and Retry-After. No numeric sustained request rate found. | | Data handling | Usage records are retained for 90 days. An explicit prompt-retention period and API training exclusion were not found in the reviewed security and privacy documents. | | Availability | The launch advertises live access. The quickstart says activation is usually immediate, while the activation page documents a capacity queue and the terms still describe early access. | | Capabilities | inference.decision | | Tags | hosted, model, closed-source, multimodal, llms-txt, usage-priced, status-page, beta | | JSON | https://www.anchorterminal.com/api/v1/tools/celeris-1-decision.json | ## Score breakdown (methodology v0.4, October 2026 research run) Assessed 2026-10-09 from public evidence against the published checklist (https://www.anchorterminal.com/benchmark/#checklist). Confidence: medium. Performance and Task success pending (no score, not in the total); the total is Σ(score × weight) ÷ 80 over the 7 assessed categories. "This run" is each category's share of the 100 points. | Category | Weight | This run | Score (0–100) | Points | | --- | --- | --- | --- | --- | | Reliability | 16% | 20 | 40 | 8.0 | | Performance | 10% | pending | pending | n/a | | Schema & documentation | 13% | 16.2 | 60 | 9.8 | | Agent ergonomics | 13% | 16.2 | 80 | 13.0 | | Security & auth | 14% | 17.5 | 45 | 7.9 | | Payments & pricing | 10% | 12.5 | 20 | 2.5 | | Task success | 10% | pending | pending | n/a | | Maintenance & community | 7% | 8.8 | 40 | 3.5 | | Transparency & trust (editorial 45, provenance 61) | 7% | 8.8 | 53 | 4.6 | | Negative events | up to −15 | up to −15 | none recorded | 0 | | **Total** | | | | **49.3 → D** | ### Why each score - Reliability 40: Component status and incident history earn 20. Only one day of history is possible for this model, so incident evidence earns 5, not a clean 90-day award. Documented 429 and backoff earn 15. Numeric request rates were not found (0). Enterprise SLA marketing gives no public commitment (0). Early-access terms prevent a general-availability award (0). The August Celeris-1 outage is a different model. - Performance: Pending. Latency is measured per call by our probes, which haven't run yet, so this run doesn't score it. Its weight is shared across the assessed categories until the first probe window closes. - Schema & documentation 60: No public machine-readable decision contract found (0). Markdown and llms.txt earn 10. Purpose and limitations earn 20, typed question variants and explicit limits 15, and examples and errors 15. Versioned paths alone do not establish a public changelog (0). Total 60. - Agent ergonomics 80: Compact probability answers earn 25. Multiple questions, bounded options and image controls earn 15 of 20; no exact token window or decision output-budget control was found. Actionable errors earn 20. Retry guidance earns 10 of 20 because duplicate inference can be billed and no idempotency contract was found. Simple required fields and documented compatible Python clients earn 10 of 15; no Celeris-owned decision SDKs in two languages found. Total 80. - Security & auth 45: Revocable bearer keys earn 20. Workspace isolation earns 10 of 20 for least privilege, but no scoped inference permissions or per-key limits were found. No decision-specific prompt-injection mitigation found (0). Usage records and trace IDs earn 15. SOC 2 appears in enterprise marketing without a reviewed report or trust evidence, so no security-programme award (0). Total 45. - Payments & pricing 20: Public per-token prices earn 20. No machine-payment protocol found (0). Card-funded activation and credit purchases prevent a card-free trial award (0). Browser signup and activation prevent autonomous onboarding credit (0). Total 20. - Task success: Pending. Task success needs the category task suites run through each tool, which haven't run yet, so this run doesn't score it. Its weight is shared across the assessed categories until then. A data provider's data-quality score is published on its listing now and becomes half of this category when it's scored. - Maintenance & community 40: The dated 8 October decision-model launch earns 30. No decision-model deprecation notice period or release history found (0). A support email is published but public response evidence was not found (0). Documented current client compatibility earns 10 of 15, not an award for Celeris-owned SDK maintenance. No owned decision package CI found (0). Total 40. - Transparency & trust 53: Closed model with service terms earns 15. Privacy and security statements explain processing and 90-day usage records, but not prompt retention or a training exclusion, earning 15 of 30. No dated deprecation policy found (0). Named subprocessors and regional-routing disclosure earn 15 of 20; a decision-specific residency guarantee was not found. Total 45. Fix list for a coding agent, everything this grade says the listing lacks, the biggest gain first (18 items): https://www.anchorterminal.com/fixes/celeris-1-decision.md (JSON https://www.anchorterminal.com/fixes/celeris-1-decision.json) ### What we couldn't check - We did not run authenticated inference or reproduce the vendor benchmark. - What are the exact decision-model context window and sustained request-rate limits? - What are prompt and response retention periods, training defaults and available data-residency commitments? - Launch copy says keys are live immediately, but activation documentation still describes a capacity queue. - The well-known security.txt and docs OpenAPI JSON addresses returned 404; no alternative public machine-readable decision contract was found. - No independent security certification evidence or public SLA terms were reviewed. ### Sources - Decision API formats, limits and examples: (seen 2026-10-09) - Decision API Markdown, fetched directly: (seen 2026-10-09) - Model routing and endpoint exclusions: (seen 2026-10-09) - Token prices and credit expiry: (seen 2026-10-09) - Workspace activation: (seen 2026-10-09) - Capacity-dependent activation queue: (seen 2026-10-09) - Key ownership and rotation: (seen 2026-10-09) - Retention and request visibility: (seen 2026-10-09) - Rate-limit scope and retries: (seen 2026-10-09) - Error codes and billing: (seen 2026-10-09) - Agent documentation index: (seen 2026-10-09) - Launch and vendor benchmark methodology: (seen 2026-10-09) - Service components: (seen 2026-10-09) - Incident history: (seen 2026-10-09) - Earlier Celeris-1 outage: (seen 2026-10-09) - Service terms and legal entity: (seen 2026-10-09) - Privacy and subprocessors: (seen 2026-10-09) - Enterprise security and SLA marketing: (seen 2026-10-09) ## Who's behind it (provenance 61/100, checked 2026-10-09) | Check | Finding | Points | | --- | --- | --- | | Legal entity named | Marqo Inc | 20/20 | | Domain age | celeris.ai, no registry record we could read | 0/15 | | Endpoint on the vendor's domain | inference.celeris.ai | 15/15 | | Terms of service | read, states 6 of the 7 things a reader expects, and has 1 clause that costs points | 7.1/10 | | Privacy policy | read, states 7 of the 8 things a reader expects | 9.3/10 | | Status page | status.celeris.ai | 10/10 | | Changelog | not found | 0/10 | | security.txt | not found | 0/10 | Terms dated 8 September 2026 and privacy notice dated 23 July 2026 name Marqo Inc operating as Celeris. No dedicated decision-model changelog or dated deprecation policy found. The public status history records an 11-hour-20-minute outage for the separate Celeris-1 model on 19 August 2026. It predates this decision model and is not counted as its outage. The standard well-known security.txt and docs OpenAPI JSON addresses returned HTTP 404. Domain registration was not checked. ### Terms and privacy, as read A reading by a fixed set of rules, each answered with the vendor's own sentence. Not legal advice. **Terms of service** (https://celeris.ai/terms), read 2026-10-09, dated 2026-09-08, states 6 of the 7 things a reader expects. - To know. Restricts benchmarking or competitive use (costs points). "Use outputs to develop competing models by systematically extracting model weights or training data." - Gives the date it was last updated. Last updated 2026-09-08. - Names the governing law or courts. The law of the State of California. - States a limit on its liability. Capped at the fees paid in the 3 months before the claim. - Says how changes to the terms are announced. Says it gives notice of a change. - Not found in the text. Refers to a service level or uptime commitment. - Also in the text (2026-10-09). Raw API access cannot be resold without written agreement. "Resell raw access to the API without our written agreement." - Also in the text (2026-10-09). An executed order form takes precedence over conflicting standard terms. "If there is a conflict between an Order Form and these terms, the Order Form prevails." **Privacy policy** (https://celeris.ai/privacy), read 2026-10-09, dated 2026-07-23, states 7 of the 8 things a reader expects. - Gives the date it was last updated. Last updated 2026-07-23. - Says how long data is kept. For as long as needed, with no period named. - Says whether personal data is sold or shared for advertising. Says it does not sell personal data. - Gives a privacy contact. support@celeris.ai. - Not found in the text. Says where data is transferred or stored. ## Live (updated 2026-10-09 11:28 UTC) - Right now: up, HTTP 404, 349 ms, checked 2026-10-09 11:28 UTC (get on `https://inference.celeris.ai/celeris-1-decision/v1/systemone`) - Uptime 24h 100.0% (21 probes) · 30 days 100.0% (21 probes) · p50 349 ms · p95 515 ms - Vendor status page: unknown, no machine-readable status found - Always current: https://www.anchorterminal.com/api/v1/live/celeris-1-decision.json ## Probe metrics Not measured yet. Our benchmark probes haven't run, so there's no availability, latency or error rate from a run and Performance is pending. Live uptime, where we poll the endpoint, is under Live and doesn't change the score. ## Prices | Item | Price | Unit | Note | | --- | --- | --- | --- | | Decision input | $0.04 | per 1M tokens | Includes cached input and image tokens. Prepaid credit expires after 30 days. | | Decision output | free | per 1M tokens | Includes optional explanations. | Across all listings: https://www.anchorterminal.com/prices/index.md ## Strengths - Typed binary, choice and score answers, with multiple questions about one state per request - Text, JSON and images through System One, with optional explanations - Decision inputs cost $0.04 per million tokens and outputs are free - Markdown documentation, error recovery guidance and revocable workspace keys ## Weaknesses - Early-access terms and capacity-dependent workspace activation - Prepaid credit expires after 30 days, with a $5 minimum card purchase - No numeric request-rate guarantee or decision-model context window found - No explicit API training exclusion or prompt-retention period found in the reviewed policies - Closed weights and no fine-tuning in this release ## Before you call it (notes for agents) 1. Use the decision model path and matching model field. This model has no chat, Responses or models endpoint. 2. Use at most 512 questions per request, or 64 with explanations. Both endpoints reject streaming. 3. Preserve x_celeris in a custom Jev SDK response model when reading explanations. The default response type drops it. 4. Retry 429 with Retry-After and backoff. Stop on 402 until workspace credit is replenished. 5. Read usage.input_tokens for cost, including images and request overhead. Track credit expiry separately from token consumption. ## Connect First request: ```bash curl https://inference.celeris.ai/celeris-1-decision/v1/systemone \ -H "Authorization: Bearer $CELERIS_API_KEY" \ -H "Content-Type: application/json" \ -d '{"model":"celeris-1-decision","state":"The customer requests a refund for a duplicate charge.","questions":{"refund":{"type":"noul","instructions":"The customer is asking for a refund."}},"x_celeris":{"explain":true}}' ``` ## Similar tools Ranked by shared capabilities, then score. Same-category tools with no shared capability key are listed last. | Tool | Grade | Score | Rank | Shared capabilities | x402 | Markdown | | --- | --- | --- | --- | --- | --- | --- | | OpenAI Decisions API | BB | 71.5 | 119 | inference.decision | no | https://www.anchorterminal.com/tools/openai-decisions-api.md | | Decider | B | 69.5 | 172 | inference.decision | no | https://www.anchorterminal.com/tools/decider.md | | Laya | B | 69.2 | 183 | inference.decision | no | https://www.anchorterminal.com/tools/convai-laya.md | | Kev | B | 67.4 | 235 | inference.decision | no | https://www.anchorterminal.com/tools/jaredpalmer-kev.md | | Vela 2.0 | B | 66.5 | 268 | inference.decision | no | https://www.anchorterminal.com/tools/vela.md | | Clef | B | 66.1 | 276 | inference.decision | no | https://www.anchorterminal.com/tools/cloudflare-clef.md | ## Panel reviews (0) Reviewed by the Anchor panel (https://www.anchorterminal.com/reviewers/index.md): . Desk reviews, written from public documentation, pricing, terms, source and status history on 1 October 2026. No calls made. For a desk review, the outcome says whether the reviewer's questions could be answered from public material: success, partial or failure. How reviews work: https://www.anchorterminal.com/reviews/how-it-works.md ## Notable - Celeris reports 80.0 per cent accuracy over 22 jev-bench datasets and 3,210 items, against 73.1 per cent for Jev 1.13.0, a 6.9 percentage-point difference. The other comparisons are Perplexity pplx-decider-v1.1-27b, Inception Mercury Decide and OpenAI gpt-6-luna (https://celeris.ai/celeris-1-decision). - The vendor reports median end-to-end latency of 67 ms for one question and 81 ms for fifty, measured from AWS us-east-1. We did not reproduce these measurements, and explanations add latency (https://celeris.ai/celeris-1-decision). - The launch identifies decision-benching harness 0cfd276. Failed requests or missing distributions score as uniform. These vendor benchmarks are not Anchor Terminal performance or task scores (https://celeris.ai/celeris-1-decision). - The legal documents identify Marqo Inc operating as Celeris. This listing covers its decision model, not the separate Celeris-1 and Magnus chat models (https://celeris.ai/terms). ## Compare - [Celeris-1 Decision vs Clef](https://www.anchorterminal.com/compare/celeris-1-decision-vs-cloudflare-clef.md): D 49.3 vs B 66.1 - [Celeris-1 Decision vs Laya](https://www.anchorterminal.com/compare/celeris-1-decision-vs-convai-laya.md): D 49.3 vs B 69.2 - [Celeris-1 Decision vs Decider](https://www.anchorterminal.com/compare/celeris-1-decision-vs-decider.md): D 49.3 vs B 69.5 - [Celeris-1 Decision vs GLiClass](https://www.anchorterminal.com/compare/celeris-1-decision-vs-gliclass.md): D 49.3 vs D 49.9 - [Celeris-1 Decision vs Kev](https://www.anchorterminal.com/compare/celeris-1-decision-vs-jaredpalmer-kev.md): D 49.3 vs B 67.4 - [Celeris-1 Decision vs Liquid d1](https://www.anchorterminal.com/compare/celeris-1-decision-vs-liquid-d1.md): D 49.3 vs E 43.5 - [Celeris-1 Decision vs OpenAI Decisions API](https://www.anchorterminal.com/compare/celeris-1-decision-vs-openai-decisions-api.md): D 49.3 vs BB 71.5 - [Celeris-1 Decision vs Strands Decider 2B](https://www.anchorterminal.com/compare/celeris-1-decision-vs-strands-decider.md): D 49.3 vs C 61.3 - [Celeris-1 Decision vs Jev](https://www.anchorterminal.com/compare/celeris-1-decision-vs-typesafe-jev.md): D 49.3 vs B 62.1 - [Celeris-1 Decision vs Vela 2.0](https://www.anchorterminal.com/compare/celeris-1-decision-vs-vela.md): D 49.3 vs B 66.5 ## Verify this listing For the vendor. The badge or a plain link to this page verifies the listing, from a page on celeris.ai or one of its subdomains. It shows the listing is the vendor's and that the vendor knows it's here, and it never changes a grade, rank or review. The vendor sends the page's address to `POST https://www.anchorterminal.com/api/v1/verify` as `{"slug": "celeris-1-decision", "url": "…"}`, or calls the `verify_listing` tool at https://www.anchorterminal.com/mcp. We fetch the page once, then again every week; two failed checks in a row and the verification lapses, and a later pass restores it. What we check: https://www.anchorterminal.com/builders/index.md#verify HTML badge: ```html Celeris-1 Decision on Anchor Terminal ``` Markdown badge, for a README: ```markdown [![Celeris-1 Decision on Anchor Terminal](https://www.anchorterminal.com/badges/celeris-1-decision.svg)](https://www.anchorterminal.com/tools/celeris-1-decision) ``` Plain link: ```html Celeris-1 Decision on Anchor Terminal ``` ## Share this listing For the vendor. Sharing assets for social media, two PNGs of 1200 × 630 that say Celeris-1 Decision is listed on Anchor Terminal, with the vendor's logo and this page's address and no grade or score. - Dark: https://www.anchorterminal.com/assets/share/celeris-1-decision-dark.png - Light: https://www.anchorterminal.com/assets/share/celeris-1-decision-light.png