Celeris-1 Decision released
Celeris · Model release
Source confirmed
Data as of October 9, 2026 · How the score is built
CurrentProprietaryUnspecified
Celeris-1 Decision
Decision readingCeleris released Celeris-1 Decision on October 8, 2026, as a hosted diffusion model for Noul, Choice, and Score questions. It accepts text, JSON, and images through a System One-compatible API, with optional one-sentence explanations. Celeris lists $0.04 per million input tokens, including cached input, and free output. Reported input usage includes the internal request representation and images.
Released Oct 8, 2026 — see all recent releases
Celeris-1 Decision will be repriced, updated or retired. Get each notice with its source and date. Follow model changes
Lineage
The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.
Decision-system
Radar
Full release historyCeleris-1 Decision release history
Radar confirmed these at the source. Use Celeris-1 Decision in your work? Explore Radar to follow supported changes and choose your alerts.
Radar
Spec sheet
Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.
- API model ID
- celeris-1-decisionCeleris Decision API documentation
- Context window
- Not publishedCeleris Decision API documentation
- Maximum output
- Not sourced yet
- Knowledge cutoff
- Not sourced yet
- Input modalities
- text, imageCeleris Decision API documentation
- Output modalities
- typed JSONCeleris Decision API documentation
- Parameters
- Not disclosed by the provider
- Availability
- Celeris Decision APICeleris-1 Decision launch report
- Cloud regions
- Not tracked yet
- Lifecycle
- Current
- API capabilities
- System One-compatible POST /v1/systemone with Noul, Choice, and Score questions over text, structured JSON, or images. Multiple mixed question types per request; optional explanations through x_celeris.explain. One JSON response, no streaming.Celeris Decision API documentation
- Prompt caching
- Published at $0.040 per million cached input tokensCeleris Decision API pricing
- Self-host
- Weights are not published
- Rate limits
- Not tracked yet
How to read this profile
The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.
We track Celeris-1 Decision, but no weighted text-model benchmark result is published on the site yet. This page shows the metadata and separate protocol evidence we can verify now; a BenchLM score will appear only if compatible public evaluations land.
Celeris-1 Decision is a proprietary model. No explicit reasoning mode is documented in this profile.
Hosted System One-compatible API for Noul, Choice, and Score questions over text, JSON, and images. Optional explanations add latency. No open weights or fine-tuning are offered in this release; an exact context limit is not published in the reviewed decision-model documentation.
Celeris reports 80.0% accuracy and 0.272 Brier on 3,210 jev-bench items, plus 76.5% accuracy and 0.079 Brier on typed decisions, using decision-benching harness 0cfd276. These provider-reported results are shown on the Decision Models page separately from the JevBench composite. No general model rank or Celeris-1 benchmark values are inherited.
The profile has no source-displayable benchmark row yet.
Model research
Celeris-reported decision results
Celeris reports 80.0% accuracy on 22 jev-bench datasets totaling 3,210 items and 76.5% on 400 typed-decisions cases with five questions each. The table preserves its October 2026 comparison, run through each provider’s decision endpoint with one scorer. Lower Brier scores mean smaller errors in the reported probabilities.
These are provider-reported results from decision-benching harness 0cfd276. This jev-bench accuracy is separate from the official JevBench composite and Perplexity’s 11-task panel. The two Brier scores use different gold labels: hard labels for jev-bench and soft distributions for typed decisions. Failed requests or missing distributions were scored as uniform. Latency was measured end to end from AWS us-east-1; optional explanations add latency. We have not rerun this evaluation, and these results stay outside general model rankings.
Last updated October 9, 2026. Runtime fields remain blank until a sourced snapshot exists.
Questions
Can I run Celeris-1 Decision locally?
Celeris-1 Decision is available through a hosted API. Celeris says this release has no open weights and no fine-tuning. System One clients can connect by changing the base URL, API key, and model name. Text, structured JSON, and images can share one request with several named questions.
Does the 80.0% result establish an overall model rank?
The 80.0% measures top-answer accuracy in Celeris’s provider-run evaluation of 3,210 jev-bench items. Its decision-benching harness differs from the JevBench composite shown on the Decision Models page. We display the report separately, leave general benchmark categories blank, and assign no overall model score or rank.