CLASSIFICATION · ROUTING · SCORING
Choose a model that selects from predefined options, scores candidates, or returns probabilities. Compare free and low-cost APIs for ticket routing, agent tool selection, and other structured decisions.
| Model / service | Access | Price & allowance | Conditions |
|---|---|---|---|
| OpenRouter · Mercury Decide | Start for free | inception/mercury-decide:free: free input and output. | 65,536-token context; 20 requests/minute and 50/day; 1,000/day after at least $10 in lifetime credit purchases. Requires an OpenRouter API key. |
| OpenRouter · Span-01 Lite | Start for free | respan/span-01-lite:free: free input and output. | Text input; 20 requests/minute and 50/day; 1,000/day after at least $10 in lifetime credit purchases. Requires an OpenRouter API key. |
| OpenCode · Jev 1.13 | Start for free | jev-1.13-free: free input and output. | Limited-time offer; end date and fixed allowance unpublished. Requires an OpenCode Console API key. |
| Venice · Jev | Start for free | Free input during the current promotion and free output; new accounts receive a daily allowance and 500 welcome credits. | API beta; no credit card. 32K state plus longest question, 64K total input. Promotion end date and daily allowance size unpublished. |
| Respan · Span-01 Lite | Start for free | span-01-free: free, with a daily cap resetting at 00:00 UTC. | Daily cap size unpublished; text input. |
| Alibaba Cloud Model Studio · Decision Model | Start for free | decision-model-preview: temporarily free in Beijing and Singapore. | 65,536-token input; 1,200 requests and 2 million tokens per minute. End date unpublished. Requires a Model Studio account and a key for the selected region. |
| Cloudflare Workers AI · Clef | Start for free | Direct Workers AI access includes 10,000 free Neurons per day, resetting at 00:00 UTC and shared across Workers AI models. | Clef-flash costs $0.038 per million input tokens with free output; paid access is available through Workers Paid or prepaid AI Gateway credits. |
| Liquid AI · d1 | Trial credits | Each new organization receives $2 credit; d1 costs $0.02 per million input tokens with free output. | Granted per new organization; credit expiry unpublished. |
| Nace · Drex | Trial credits | Receive $25 at signup, valid for 3 months; $5 monthly starting one month after signup, expiring at the next renewal. | Once per mailbox; disposable email excluded. Up to 3 eligible accounts per business domain per day. Without a card, calls stop when credit runs out. |
| Liquid AI · Open d1 | Open weights & local use | LFM Open License: free commercial use below $10M annual revenue; research, education and nonprofit use is free. | Larger businesses need a commercial production license; supply your own compute. |
| TypeSafe AI · Jev | Lower-cost APIs | Input costs $0.042 per million tokens. | Requires a TypeSafe API key. |
| Respan · Span-01 | Lower-cost APIs | span-01-pro: $0.02 per million input tokens with free output. | Requires Respan credits; billed input includes the trace and all behavior definitions. |
| Vercel AI Gateway | Lower-cost APIs | Jev: $0.042 per million input tokens; Liquid d1: $0.02 per million input tokens. Output is free for both. | Jev, Jev 1.14 and Liquid d1 require purchased AI Gateway credits; these models are excluded from the $5/month free tier. |
| Respan · Span-02 | Lower-cost APIs | $0.01 per million input tokens; output and thought tokens are free. | Requires Respan credits; up to 4 images per request. |
| Respan · Span-02 Open Weights | Open weights & local use | Free weights and inference engine under Apache-2.0. | Requires a Blackwell RTX 5090 or RTX PRO 6000; default settings use about 28 GB of GPU memory. Use the included Rust/CUDA engine and supply your own compute. |
Test classification, routing, and scoring with a free API.
Classify, score and evaluate text or images.
Score custom conversation behaviors as present, absent or not observable.
Call Jev for classification, scoring and yes/no decisions with one API key.
Turn tickets and business records into probabilities, choices and scores.
Score agent conversations against custom behavior definitions, returning present, absent and not-observable probabilities.
Classify, score and evaluate business state through the System One API.
Run the Clef family through Workers or REST for classification, scoring and multimodal decisions.
Claim eligible credits to test models on your own tasks.
Classify, route and score text or images with structured decisions.
Classify, evaluate yes/no questions and score state through a System One-compatible API.
Download a model and run it on your hardware or in the cloud.
Run d1-3B and d1-omni-600M on your own devices or servers.
Run Span-02 on your own GPU for text and image decisions.
Compare billing units and access conditions for ongoing API use.
Call Jev directly to turn shared state into probabilistic choices, scores and yes/no decisions.
Evaluate custom conversation behaviors for agent assessment and routing.
Call Jev, Liquid d1 and other decision models through one gateway with AI SDK integration and usage controls.
Return choices, scores or numbers from text, JSON and images with a configurable reasoning deadline.
New credits, free access campaigns, and previous offers for these models.
2 deals · page 1 of 1
Jev is free for a limited time on OpenCode Zen. Mercury Decide and Span-01 Lite offer free API access, while new Liquid organizations receive $2 in trial credits. Kev and Laya can be hosted on your own hardware.
Alibaba Cloud Model Studio's new decision model, decision-model-preview, is free to call for a limited time in both the Singapore and China (Beijing) regions. One request runs classification, yes/no checks, and scoring in parallel and returns probabilities with confidence scores.
