{ "meta": { "attribution": "Anchor Terminal (https://www.anchorterminal.com)", "docs": "https://www.anchorterminal.com/docs/", "generatedAt": "2026-10-09", "license": "CC-BY-4.0", "method": "https://www.anchorterminal.com/benchmark/", "methodology": "0.4", "openapi": "https://www.anchorterminal.com/openapi.json", "preview": false, "run": "2026-10-01", "runLabel": "October 2026 research run" }, "tool": { "slug": "decider", "name": "Decider", "vendor": "Mark Marosi (Mapika)", "vendorUrl": "https://github.com/Mapika", "kind": "model", "category": "decision-models", "summary": "Decider is a family of open-weight decision models by Mark Marosi (Mapika), from 0.8B to 35B parameters under Apache-2.0. It answers typed yes or no, choice and score questions with probabilities, and runs locally from the `decider-ai` Python package.", "url": "https://www.anchorterminal.com/tools/decider", "markdownUrl": "https://www.anchorterminal.com/tools/decider.md", "slimMarkdownUrl": "https://www.anchorterminal.com/tools/decider.min.md", "jsonUrl": "https://www.anchorterminal.com/api/v1/tools/decider.json", "repo": "https://github.com/Mapika/decider", "license": "Apache-2.0 (code and weights)", "transports": [ "http" ], "packages": [ { "registry": "pypi", "name": "decider-ai" } ], "auth": "none", "authNotes": "No account. `decider.serve` has no authentication option. `scripts/serve.sh` binds to 127.0.0.1 since 1.7.1, and `DECIDER_HOST=0.0.0.0` opens it to the network. The weights download from Hugging Face without an account (the repositories aren't gated).", "pricing": "free", "pricingNotes": "Free and open source, with nothing to buy. You pay for your own hardware. The README puts the 2B at about 4 GB of GPU memory, the 4B at 8.4 GB and the 35B at 65 GB in bf16, with GGUF files of 1.3 GB and 2.7 GB for CPU (https://github.com/Mapika/decider). No hosted API was found.", "priceSummary": "Free ยท OSS", "where": "local", "x402": { "level": "no", "evidence": "No x402, MPP or L402. Decider is software you run, and its server has no payment route (checked 2026-10-08).", "endpoints": [] }, "toolCount": null, "popularity": { "githubStars": 1100, "npmWeekly": null, "pypiWeekly": 2680, "asOf": "2026-10-08" }, "docsUrl": "https://github.com/Mapika/decider#readme", "capabilities": [ "inference.decision" ], "tags": [ "model", "open-source", "open-weights", "self-hosted", "local", "free", "python" ], "lastRelease": "2026-10-07", "graded": true, "anchor": { "graded": true, "score": 69.5, "grade": "B", "agentReady": false, "rank": 172, "ranked": true, "rankOf": 842, "categoryRank": 2, "methodology": "0.4", "run": "2026-10-01", "scores": { "ergonomics": 80, "maintenance": 87, "payments": 60, "reliability": 90, "schema": 78, "security": 38, "transparency": 46 }, "pending": [ "performance", "tasks" ], "breakdown": [ { "key": "reliability", "name": "Reliability", "weight": 16, "effectiveWeight": 20, "score": 90, "points": 18, "reason": "Scored on the local-package checklist, since Decider is open weights the owner runs, as for Kev and Strands Decider. `pip install decider-ai` from PyPI, 1.9.0 of 7 October 2026, with Python 3.11 or later stated and extras for serving, GGUF and Apple silicon (20). Public CI on GitHub Actions runs 26 test files in two jobs, without torch and with CPU torch, and the five latest runs on main had passed on 8 October. No test loads model weights, and the changelog reports 294 passing (22 of 25). 3 of 16 issues are open, a feature request, a Windows on ARM report with six comments and fixes shipped in 1.7.1, and one unsolicited promotion. Reported crashes (#5, #8, #21) were fixed in a release within days (23 of 25). `docs/CHANGELOG.md` has a dated entry for every release and tags match PyPI. The 1.3.0 minor release changed what `confidence` means, called out with a migration note (12 of 15). 1.9.0, though 1.0.0 is from 22 September 2026 and 21 releases followed in 16 days (13 of 15)." }, { "key": "performance", "name": "Performance", "weight": 10, "effectiveWeight": 0, "pending": true, "points": 0, "reason": "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." }, { "key": "schema", "name": "Schema \u0026 documentation", "weight": 13, "effectiveWeight": 16.25, "score": 78, "points": 12.68, "reason": "Read for a model you serve yourself. The server is FastAPI with Pydantic request models, but `questions` is an untyped dict and `state` any JSON value, with the question rules enforced in `decider/systemone.py`. It follows TypeSafe's wire format and publishes no spec file of its own (15 of 25). No llms.txt. The README, five model cards and six docs pages are Markdown in the repository (5 of 10). The README says which model to use for which hardware and has a limits section with measured figures for each failure (19 of 20). Three question types, 2 to 255 options a choice and 2 to 10 levels a score, checked with a 422. State is free-form by design (11 of 15). Python examples with sample output and three example programs. `docs/SERVING.md` gives the 413, 503 and 422 bodies. No curl example and no error table (13 of 15). Release tags, Hub tags for earlier weights and a dated changelog (15)." }, { "key": "ergonomics", "name": "Agent ergonomics", "weight": 13, "effectiveWeight": 16.25, "score": 80, "points": 13, "reason": "Read as an API an agent calls for a decision, as for the other decision models. Answers are a probability per option, and a long state is read once with each question scored from the shared prefix. Context is 32k tokens, and a state over 32,768 tokens is truncated without an error (19 of 25). The caller sets the questions, any number a request up to 1,024 rows, with `independent` and a plain `/decide` form. No batch-of-states route (17 of 20). 422, 413 and 503 return a message that names the limit and its variable, documented in `docs/SERVING.md` (17 of 20). Calls are stateless and safe to retry, and the 503 body says to retry later. No Retry-After header (16 of 20). One pip install and a Python class with a default device order. The README says TypeSafe's SDKs work unchanged. Python only, and `scripts/serve.sh` comes from a clone (11 of 15)." }, { "key": "security", "name": "Security \u0026 auth", "weight": 14, "effectiveWeight": 17.5, "score": 38, "points": 6.65, "reason": "Read as software you run. No account. The server has no authentication option, binds to 127.0.0.1 since 1.7.1 and the script comment says so (8 of 30). A decision model has no write actions, so there's nothing to approve (15 of 20). Nothing documents how hostile text in the state can move an answer. The README says rules written into a question aren't followed at this size (5 of 15). `/stats` counts requests, errors and rejections, and each response carries token usage. No request log (5 of 15). No SECURITY.md, disclosure policy or advisories found. Weights ship as safetensors, PyPI uploads use trusted publishing with no token in the repository, and workflow actions are pinned by tag, not commit (5 of 20)." }, { "key": "payments", "name": "Payments \u0026 pricing", "weight": 10, "effectiveWeight": 12.5, "score": 60, "points": 7.5, "reason": "Free Apache-2.0 software with nothing to buy, so 20 + 20 + 20 for pricing, free use and no sign-up, by the self-hosted rule. No payment protocol (0). No hosted API was found, so there's no hosted option to grade." }, { "key": "tasks", "name": "Task success", "weight": 10, "effectiveWeight": 0, "pending": true, "points": 0, "reason": "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." }, { "key": "maintenance", "name": "Maintenance \u0026 community", "weight": 7, "effectiveWeight": 8.75, "score": 87, "points": 7.61, "reason": "Read for an open-weight model. `decider-ai` 1.9.0 and the decider-31b weights on 7 October 2026 (30). 21 package releases since 22 September 2026 (20). 13 of 16 issues closed, with fixes released within days of reports #5, #8, #18 and #21, and outside pull requests merged (#2, #11, #13). One maintainer wrote 95 of the 98 commits under two names (21 of 25). A Python package on PyPI from the author, and TypeSafe's SDKs for the HTTP route. Not an MCP server, so no registry entry applies (10 of 15). CI passes on main. Dependencies have no upper bounds or lockfile, and the vLLM path needs its own environment pinned to vLLM 0.29.0 (6 of 10)." }, { "key": "transparency", "name": "Transparency \u0026 trust", "weight": 7, "effectiveWeight": 8.75, "score": 46, "points": 4.03, "note": "editorial 64, provenance 27", "reason": "Apache-2.0 for the code and model repositories, with the training code, data builders, teacher data and per-stage measurements published. The RL stage and the mixture-v2 builders aren't in the package, and the README says reproduction isn't byte-identical (28 of 30). Self-hosted, so inputs stay on the operator's hardware by construction, but we found no statement saying so. Training draws on about 95 public datasets whose licences we didn't review (15 of 30). The 1.0.x server was kept for one release as `decider.serve_v1` with notice in the changelog, and earlier weights stay under Hub tags. No deprecation policy (11 of 20). No telemetry code found in the package and no statement either way. Weights download through the Hugging Face Hub client (10 of 20)." } ], "assessment": { "date": "2026-10-08", "basis": "public evidence", "confidence": "medium", "notes": { "ergonomics": "Read as an API an agent calls for a decision, as for the other decision models. Answers are a probability per option, and a long state is read once with each question scored from the shared prefix. Context is 32k tokens, and a state over 32,768 tokens is truncated without an error (19 of 25). The caller sets the questions, any number a request up to 1,024 rows, with `independent` and a plain `/decide` form. No batch-of-states route (17 of 20). 422, 413 and 503 return a message that names the limit and its variable, documented in `docs/SERVING.md` (17 of 20). Calls are stateless and safe to retry, and the 503 body says to retry later. No Retry-After header (16 of 20). One pip install and a Python class with a default device order. The README says TypeSafe's SDKs work unchanged. Python only, and `scripts/serve.sh` comes from a clone (11 of 15).", "maintenance": "Read for an open-weight model. `decider-ai` 1.9.0 and the decider-31b weights on 7 October 2026 (30). 21 package releases since 22 September 2026 (20). 13 of 16 issues closed, with fixes released within days of reports #5, #8, #18 and #21, and outside pull requests merged (#2, #11, #13). One maintainer wrote 95 of the 98 commits under two names (21 of 25). A Python package on PyPI from the author, and TypeSafe's SDKs for the HTTP route. Not an MCP server, so no registry entry applies (10 of 15). CI passes on main. Dependencies have no upper bounds or lockfile, and the vLLM path needs its own environment pinned to vLLM 0.29.0 (6 of 10).", "payments": "Free Apache-2.0 software with nothing to buy, so 20 + 20 + 20 for pricing, free use and no sign-up, by the self-hosted rule. No payment protocol (0). No hosted API was found, so there's no hosted option to grade.", "reliability": "Scored on the local-package checklist, since Decider is open weights the owner runs, as for Kev and Strands Decider. `pip install decider-ai` from PyPI, 1.9.0 of 7 October 2026, with Python 3.11 or later stated and extras for serving, GGUF and Apple silicon (20). Public CI on GitHub Actions runs 26 test files in two jobs, without torch and with CPU torch, and the five latest runs on main had passed on 8 October. No test loads model weights, and the changelog reports 294 passing (22 of 25). 3 of 16 issues are open, a feature request, a Windows on ARM report with six comments and fixes shipped in 1.7.1, and one unsolicited promotion. Reported crashes (#5, #8, #21) were fixed in a release within days (23 of 25). `docs/CHANGELOG.md` has a dated entry for every release and tags match PyPI. The 1.3.0 minor release changed what `confidence` means, called out with a migration note (12 of 15). 1.9.0, though 1.0.0 is from 22 September 2026 and 21 releases followed in 16 days (13 of 15).", "schema": "Read for a model you serve yourself. The server is FastAPI with Pydantic request models, but `questions` is an untyped dict and `state` any JSON value, with the question rules enforced in `decider/systemone.py`. It follows TypeSafe's wire format and publishes no spec file of its own (15 of 25). No llms.txt. The README, five model cards and six docs pages are Markdown in the repository (5 of 10). The README says which model to use for which hardware and has a limits section with measured figures for each failure (19 of 20). Three question types, 2 to 255 options a choice and 2 to 10 levels a score, checked with a 422. State is free-form by design (11 of 15). Python examples with sample output and three example programs. `docs/SERVING.md` gives the 413, 503 and 422 bodies. No curl example and no error table (13 of 15). Release tags, Hub tags for earlier weights and a dated changelog (15).", "security": "Read as software you run. No account. The server has no authentication option, binds to 127.0.0.1 since 1.7.1 and the script comment says so (8 of 30). A decision model has no write actions, so there's nothing to approve (15 of 20). Nothing documents how hostile text in the state can move an answer. The README says rules written into a question aren't followed at this size (5 of 15). `/stats` counts requests, errors and rejections, and each response carries token usage. No request log (5 of 15). No SECURITY.md, disclosure policy or advisories found. Weights ship as safetensors, PyPI uploads use trusted publishing with no token in the repository, and workflow actions are pinned by tag, not commit (5 of 20).", "transparency": "Apache-2.0 for the code and model repositories, with the training code, data builders, teacher data and per-stage measurements published. The RL stage and the mixture-v2 builders aren't in the package, and the README says reproduction isn't byte-identical (28 of 30). Self-hosted, so inputs stay on the operator's hardware by construction, but we found no statement saying so. Training draws on about 95 public datasets whose licences we didn't review (15 of 30). The 1.0.x server was kept for one release as `decider.serve_v1` with notice in the changelog, and earlier weights stay under Hub tags. No deprecation policy (11 of 20). No telemetry code found in the package and no statement either way. Weights download through the Hugging Face Hub client (10 of 20)." }, "sources": [ { "what": "repository at commit e50e549, README, model cards, server code, tests, workflows and scripts (cloned)", "url": "https://github.com/Mapika/decider", "seen": "2026-10-08" }, { "what": "changelog", "url": "https://github.com/Mapika/decider/blob/main/docs/CHANGELOG.md", "seen": "2026-10-08" }, { "what": "serving design, defaults, limits and error responses", "url": "https://github.com/Mapika/decider/blob/main/docs/SERVING.md", "seen": "2026-10-08" }, { "what": "PyPI package metadata and release history", "url": "https://pypi.org/pypi/decider-ai/json", "seen": "2026-10-08" }, { "what": "PyPI download counts", "url": "https://pypistats.org/api/packages/decider-ai/recent", "seen": "2026-10-08" }, { "what": "Hugging Face models by Mapika", "url": "https://huggingface.co/api/models?author=Mapika", "seen": "2026-10-08" }, { "what": "decider-2b repository metadata and config", "url": "https://huggingface.co/api/models/Mapika/decider-2b", "seen": "2026-10-08" }, { "what": "decider-4b model card", "url": "https://huggingface.co/Mapika/decider-4b", "seen": "2026-10-08" }, { "what": "issue list", "url": "https://github.com/Mapika/decider/issues?q=is%3Aissue", "seen": "2026-10-08" }, { "what": "CI runs of the tests workflow", "url": "https://github.com/Mapika/decider/actions/workflows/tests.yml", "seen": "2026-10-08" } ], "openQuestions": [ "unchecked: the GitHub API answered with a rate limit for our address, so the star count is the rounded 1.1k from the repository page (47 forks) and the pull request list was not read", "unchecked: the licences of the roughly 95 training datasets, and the licence terms of the Gemma-4 bases under decider-12b and decider-31b, whose cards we didn't read", "The lead said the model card states it is unrelated to Strands Decider 2B. No such statement is in the repository or the decider-4b card. The independence statement there concerns TypeSafe AI", "The lead named three models. The family has more, among them decider-0.8b, decider-12b and decider-31b", "The leaderboard positions and all accuracy, calibration and speed figures are the author's or third parties' as cited by the author. We haven't run them", "No legal entity. The author is an individual, and commits appear under the names Mapika and Mark Marosi", "The slug `decider` sits beside `strands-decider`, an unrelated product with the same name" ] }, "negative": 0, "verdict": "An Apache-2.0 decision model family with a dated changelog, passing CI, 21 package releases since 22 September 2026 and model cards that list measured regressions. One person maintains it, the local server has no authentication option, states over 32,768 tokens are cut without an error, and no security policy is published.", "bestFor": "Local classification, routing, triage and checks where a team wants open weights in several sizes and a Jev-shaped route.", "strengths": [ "Apache-2.0 code and weights, with the training code, data builders and per-version measurements in the repository", "Sizes from 0.8B to 35B parameters, with GGUF files for CPU and builds for CUDA, Apple silicon and vLLM", "`POST /v1/systemone` follows TypeSafe's wire format, and the README says TypeSafe's SDKs work with `TYPESAFE_BASE_URL` set to the local server", "Dated changelog entries for all 21 `decider-ai` releases from 1.0.0 (22 September 2026) to 1.9.0 (7 October 2026)", "Size limits return 413, overload 503 and invalid questions 422, each with a message documented in `docs/SERVING.md`" ], "weaknesses": [ "The local server has no authentication option. It binds to 127.0.0.1 since 1.7.1", "States over 32,768 tokens are truncated (`DECIDER_MAX_STATE_TOKENS`) without an error", "No SECURITY.md, disclosure policy or security contact was found in the repository", "One maintainer wrote 95 of 98 commits, and the project is three weeks old", "The README says calibration on hard items is weak, the models are English only, and one pass can't do multi-step arithmetic", "No hosted API, and the HTTP server doesn't serve the GGUF files" ], "agentNotes": [ "Pin weights by Hub tag (`v10`, `v2`) when results must repeat. The `main` branch of each model repository changes with new versions", "Read `x_p_max` for the top probability. Since 1.3.0 `confidence` on choice and score answers follows TypeSafe's rescaled definition, not the top probability", "Keep the server on 127.0.0.1 or put an authenticating proxy in front. It has no key option", "Count state tokens before sending. Over 32,768 the state is cut silently, and `/decide` caps context at 1,536 tokens", "Split multi-step arithmetic or multi-hop judgements into several questions, and don't write long rules into a question. The README says both fail" ], "metrics": { "kind": "local", "measured": false }, "reviewCount": 0, "avgRating": 0, "history": [ { "basis": "public evidence", "confidence": "medium", "grade": "B", "methodology": "0.4", "pending": [ "performance", "tasks" ], "run": "2026-10-01", "runLabel": "October 2026 research run", "score": 69.5 } ], "editorialScores": { "ergonomics": 80, "maintenance": 87, "payments": 60, "reliability": 90, "schema": 78, "security": 38, "transparency": 64 }, "provenanceScore": 27 }, "connect": { "install": "pip install decider-ai" }, "letme": { "capability": "https://letme.dev/inference.decision", "tool": "https://letme.dev/decider" }, "notable": [ "Described as an independent open reproduction of the System One model class (TypeSafe's Jev), not affiliated with TypeSafe AI, with nothing distilled from Jev, per the README (https://github.com/Mapika/decider)", "The models on Hugging Face under Mapika are decider-0.8b, decider-2b (v11), decider-4b (v2.1), decider-12b (v2), decider-31b, decider-35b-a3b, an NVFP4 build, a vision variant, GGUF files for the 2B and 4B, and two `decider-chat` configurations of stock models (https://huggingface.co/Mapika)", "`decider-ai` 1.9.0 was published on PyPI on 7 October 2026, the 21st release since 1.0.0 on 22 September 2026 (https://pypi.org/pypi/decider-ai/json)", "The README cites two third-party leaderboards read on 29 September 2026, with decider-4b v2 7th of 99 on JevBench v1.5.2, and decider-chat-gemma4-31b 2nd of 70 on the Decision Index v0.2.1. These are the author's citations, not our measurements (https://github.com/Mapika/decider#standing)", "The README's limits section lists weak calibration on hard items (top-label ECE 0.18 for decider-2b v11 on JevBench public hard items), English only, and named regressions between versions (https://github.com/Mapika/decider#limits-stated-plainly)", "`decider.serve` has routes `POST /v1/systemone`, `POST /decide`, `GET /v1/models`, `GET /health` and `GET /stats`, with defaults and limits in `docs/SERVING.md` (https://github.com/Mapika/decider/blob/main/docs/SERVING.md)", "Hugging Face showed 330,133 downloads and 102 likes for Mapika/decider-2b on 8 October 2026 (https://huggingface.co/api/models/Mapika/decider-2b)", "No relation to the listed Strands Decider 2B from AWS, which shares the name. Neither the repository nor the decider-4b card mentions it" ], "area": "models", "details": [ { "label": "Models", "value": "decider-0.8b, decider-2b v11 (1.9B), decider-4b v2.1 (4.2B) and decider-35b-a3b v1 (34.7B, 3B active) on Qwen3.5 bases. decider-12b v2 and decider-31b on Gemma-4 instruct bases. Two `decider-chat` repositories are stock models with a config" }, { "label": "Licence", "value": "Apache-2.0 for the code and the model repositories. Training uses about 95 public datasets under their own licences, which we didn't review" }, { "label": "Question types", "value": "noul (probability of yes), choice (2 to 255 options) and score (2 to 10 levels), any number a request, up to 1,024 scoring rows" }, { "label": "Context", "value": "32k tokens per the README. The server cuts states at 32,768 tokens (`DECIDER_MAX_STATE_TOKENS`), and `/decide` caps context at 1,536" }, { "label": "Input", "value": "State as text or JSON. Images only with the separate decider-2b-vision model, which the README says is on older text weights" }, { "label": "Hardware", "value": "CUDA (about 4 GB for the 2B, 8.4 GB for the 4B, 65 GB for the 35B in bf16), Apple silicon (MPS, optional MLX kernel), CPU, and llama.cpp for the GGUF files. decider-31b needs vLLM and NVFP4" }, { "label": "Hosted option", "value": "None found" }, { "label": "Training", "value": "`scripts/train.sh full` reproduces the supervised stages of the 2B, 5.3 hours on a GH200 per the README. The RL stage and the mixture-v2 builders aren't in the package" } ], "provenance": { "legalEntity": "", "domain": "github.com/Mapika", "domainRegistered": "", "endpointOnVendorDomain": null, "terms": "", "privacy": "", "statusPage": "", "changelog": "https://github.com/Mapika/decider/blob/main/docs/CHANGELOG.md", "securityTxt": "none", "checked": "2026-10-08", "notes": [ "An individual's open-source project under Apache-2.0. The pyproject and the README citation name Mark Marosi as author, and no company is named.", "No vendor domain. The code is at github.com/Mapika/decider and the weights at huggingface.co/Mapika, so the domain line names the GitHub account and scores no domain age.", "Software you run, so there's no hosted endpoint, service terms or privacy policy. The author publishes none, and the Apache-2.0 licence stands in for terms.", "The changelog is `docs/CHANGELOG.md`, with a dated entry for each package release and model update. Release tags v1.0.2 to v1.9.0 match the PyPI versions." ], "score": 27, "checks": [ { "check": "Legal entity named", "value": "not found", "points": 0, "max": 20, "state": "no" }, { "check": "Domain age", "value": "github.com/Mapika, no registry record we could read", "points": 0, "max": 15, "state": "no" }, { "check": "Endpoint on the vendor's domain", "value": "no hosted endpoint", "points": 0, "max": 0, "state": "na" }, { "check": "Terms of service", "value": "nothing hosted, so the Apache-2.0 (code and weights) licence stands in", "points": 10, "max": 10, "state": "ok" }, { "check": "Privacy policy", "value": "nothing hosted, not scored", "points": 0, "max": 0, "state": "na" }, { "check": "Status page", "value": "not found", "points": 0, "max": 10, "state": "no" }, { "check": "Changelog", "value": "published", "points": 10, "max": 10, "state": "ok" }, { "check": "security.txt", "value": "not found", "points": 0, "max": 10, "state": "no" } ] }, "pageJsonUrl": "https://www.anchorterminal.com/tools/decider.json" } }