Skip to content
ReviewRamble

The weekly brief

AI, reviewed, in your inbox.

  • New reviews of tools, models, MCP servers and repos, with the verdict up front
  • Re-checked prices and ratings when something changes
  • New analysis pieces as they are published

Models, agents, MCP servers and the tools actually worth your time.

One email a week. Unsubscribe with the link in any brief. How we use your email

The best AI decision models in 2026

The best AI decision models right now are Perplexity Decider (pplx-decider), Qwen3Guard and Llama Prompt Guard 2, out of 18 we have reviewed. Each is ranked by our rating, with its price, our verdict and who it suits.

By ReviewRamble editors · Updated Nov 9, 2026 · How we review

How we ranked: by our rating out of 5, from reviews built on each product's own pricing, documentation and policies plus independent benchmarks where they exist, each checked on the date shown in the review. We take no affiliate commission.

  1. 1

    A decision model you can download, read images with and run for under half the price of Jev, scoring well above Jev on tests nobody can train on.

    Best for: Teams that want strong typed decisions over text and images, either through a cheap API or on their own GPUs.

    Price: API: pplx-decider-v1.2 $0.02 and Flash $0.01 per 1M input tokens; output free, no per-request fee. Weights free under Apache 2.0.

    Read our Perplexity Decider (pplx-decider) review
    4.5/5
  2. 2

    The open guard model to beat for multilingual moderation, with a controversial label that saves you arguing over borderline cases and a streaming mode that checks answers while they are written.

    Best for: Teams moderating chat or model output in many languages who want an open model with a tunable middle ground between safe and unsafe.

    Price: Free open weights under Apache 2.0, ungated on Hugging Face, in 0.6B, 4B and 8B sizes; self-hosted. Alibaba Cloud also offers a hosted AI Guardrails service powered by it.

    Read our Qwen3Guard review
    4.3/5
  3. 3

    A cheap first filter for injected instructions that every agent pipeline can afford, as long as nobody mistakes it for the whole defence.

    Best for: Teams that want a fast, self-hosted check for injected instructions on user input, retrieved documents and tool results.

    Price: Free open weights under the Llama 4 Community License, gated on Hugging Face; self-hosted only, in 86M and 22M sizes.

    Read our Llama Prompt Guard 2 review
    4.2/5
  4. 4

    A small, open moderation model that adapts to your own rules without retraining, matching a 20B rival on Mistral's tests; each extra policy costs another pass.

    Best for: Teams building their own moderation or guardrail layer who need to enforce custom policies on text and images without retraining a model.

    Price: Free open weights under Apache 2.0, gated on Hugging Face; self-hosted only, on a single 16 GB GPU in BF16.

    Read our Shieldstral 1.0 3B review
    4.2/5
  5. 5

    d3-turbo earns its place when one input faces many questions, and costs more than d3-flash when it faces only one.

    Best for: Teams with GPUs who ask many yes/no, choice or location questions of the same text, image or video and want open weights.

    Price: Free weights under Apache 2.0. No paid hosted API; a free public API with logging and no SLA.

    Read our d3-turbo (vLLM Semantic Router) review
    4.1/5
  6. 6

    Strong, open and easy to serve on vLLM, but nine days after launch its own d3-turbo made it the version to migrate from.

    Best for: Teams already running vLLM who want an open decision model for text, image or video checks on their own GPUs and have validated against this checkpoint.

    Price: Free weights under Apache 2.0. No paid hosted API; the free public demo now serves d3-turbo.

    Read our d3 (vLLM Semantic Router) review
    4.0/5
  7. 7

    The right tool when your moderation policy changes faster than you can label training data, and the wrong one for screening every message at low latency.

    Best for: Trust and safety teams who write and revise their own policies and need an explanation attached to each decision.

    Price: Free open weights under Apache 2.0 on Hugging Face; self-hosted. The 20b fits on a 16 GB GPU; the 120b scores higher on OpenAI's multi-policy test.

    Read our gpt-oss-safeguard review
    4.0/5
  8. 8

    An open judge for English-language apps that covers harm, grounding and tool calls in one model, though at 8B it is a spot-checker rather than an inline filter for every message.

    Best for: English-language teams that want one open judge for safety, retrieval grounding and agent tool calls, run on sampled or high-risk traffic.

    Price: Free open weights under Apache 2.0, ungated on Hugging Face; self-hosted only, no hosted inference provider listed.

    Read our Granite Guardian 4.1 8B review
    4.0/5
  9. 9

    Microsoft-Decision-1 makes decision models a default Azure option at Jev's price, but its benchmark lead is self-reported and open rivals do the same job for less.

    Best for: Azure teams that need cheap, fast classification, routing or gating decisions under Microsoft's data terms, and will validate thresholds on their own labelled data.

    Price: $0.042 per 1M input tokens; output tokens free. Billed through an Azure subscription on Foundry, or through OpenRouter.

    Read our Microsoft-Decision-1 review
    3.9/5
  10. 10

    Jev turned typed decisions into a product, and open models have since passed it on both score and price.

    Best for: Teams that want a hosted, low-cost API for text routing, triage and yes/no gates without running their own GPUs.

    Price: $0.042 per 1M input tokens ($42 per billion); output tokens are free. Same price at long context. Enterprise limits through sales.

    Read our Jev review
    3.8/5
  11. 11

    DecisionNode answers more kinds of question than Jev at the same price, but every number behind it is Bynn's own and preliminary, and your requests train its models unless you opt out.

    Best for: Teams that need typed decisions about images and video, including coordinates, behind one API, and are willing to switch off training and run their own evaluation first.

    Price: DecisionNode $0.042 and Flash $0.035 per 1M input tokens; images and video $0.09 and $0.08 per 1M; output free; batch 30% and 50% off. Free daily tier of 50 Flash and 20 DecisionNode requests. Edge weights free under AGPL-3.0; commercial licence by request.

    Read our DecisionNode review
    3.7/5
  12. 12

    The prices are real and the classifier is good, but the launch numbers outran the evidence and the Jev compatibility costs more than it looks.

    Best for: Teams paying flagship-model prices for routing, tagging or field extraction who want a cheap hosted replacement and will test it on their own labelled data first.

    Price: $0.05 per 1M text input tokens, $0.15 per 1M image input tokens, output free; 1M free tokens for new accounts. OpenRouter: $0.04 with a 20% discount. Enterprise: dedicated instance or self-host licence, by quote.

    Read our ScaleDown Constrained Decoding Models (Classify, Extract, Summarize) review
    3.7/5
  13. 13

    Clef makes sense inside a Cloudflare Worker and almost nowhere else: Perplexity's Decider scores higher, reads images too and costs a twelfth as much.

    Best for: Teams already running on Cloudflare Workers who want typed decisions about images next to their code and value data terms over leaderboard rank.

    Price: Workers AI: $0.24 per 1M input tokens (images count as input, at most 1,024 tokens each); output free. Weights free under Apache 2.0.

    Read our Clef review
    3.6/5
  14. 14

    GLiNER2.5-Decide is a fast, honest label classifier that performs well on Fastino's tasks and far behind on everyone else's.

    Best for: Engineers who need cheap, CPU-only intent routing, moderation or spam decisions and can fine-tune on their own labels.

    Price: Free weights and library under Apache 2.0. Fastino's hosted fine-tuning agent is a separate product.

    Read our GLiNER2.5-Decide review
    3.6/5
  15. 15

    Guard 4 is the simple choice for screening text and images against a standard harm list, provided you accept that it misses about a third of unsafe English replies on Meta's own test.

    Best for: Teams outside the EU that need one self-hosted check for text and images against the MLCommons harm categories, ahead of or behind an existing chat model.

    Price: Free weights under the Llama 4 Community License (gated on Hugging Face). Hosted on DeepInfra via OpenRouter at $0.18 input and $0.18 output per 1M tokens ($0.144 on the flex tier).

    Read our Llama Guard 4 review
    3.6/5
  16. 16

    Honest numbers and good calibration under Apache 2.0, but at 40.75 on the community index it trails the team's own d3-mini by nine points.

    Best for: Engineers who need a small, self-hosted, permissively licensed text classifier whose probabilities can be trusted for thresholds.

    Price: Free weights under Apache 2.0; no hosted endpoint for this model.

    Read our Decision-2.0-Nox-4B (vLLM Semantic Router) review
    3.4/5
  17. 17

    Very fast and genuinely open, but close to guessing on hard decisions out of the box: a base to fine-tune, not a Jev replacement.

    Best for: Teams with their own labelled decision data who want a tiny, self-hosted model to fine-tune for fast routing or guardrails.

    Price: Free weights and Python package under Apache 2.0. No hosted paid API; you run it yourself.

    Read our Laya review
    3.4/5
  18. 18

    A curiosity rather than a component: tiny and fast, but it ranks 228th of 247 on the community decision index and its own maker has already replaced it.

    Best for: Researchers and hobbyists testing how far an audio-capable decision model can shrink, not teams choosing a production classifier.

    Price: Free weights under the LFM Open License v1.0; commercial use needs a licence from Liquid AI at $10 million or more in annual revenue. No hosted endpoint for this checkpoint.

    Read our d1-omni-600M review
    3.0/5

Sources

The main official source behind each ranked review; every review lists its full set.

More guides