anytimelime — the free AI field report

2026-09-14 · generated by a free model interviewing free models · anytimelime, formerly Project42 · get the engine →

What's free right now

8 free models made a real tool call today, out of 29 discovered, and the morning diff tells the story: north-mini-code, deepseek-ai/deepseek-v4-flash-0731, openai, qwen3.8-27b, nvidia/nemotron-3-super-120b-a12b joined the tool-call club today, and deepseek-ai/DeepSeek-V4-Flash-0731, inclusionai/ling-3.0-flash-sante:free, inclusionai/ling-3.0-flash-vl:free lost theirs since the 2026-09-13 report. The table below is measured, not marketed — each model got a real request with a real tool schema, timed from request to response. It says nothing a catalog can't; what it doesn't say is how long any of this survives.

How to use them with AnytimeLime

Clone the engine from github.com/fusionpoweredsoftware/anytimelime, set LIME_KEY to your OpenRouter key, and run ./lime.sh. It fetches today's verified roster and rotates across the passing models so no single free tier carries all your traffic. Your key stays on your own machine — the website never asks for it.

The churn

Free rosters change weekly — models die, rate-limit, or resurrect without notice. Last week's config is this week's outage. Re-squeeze before you depend on it.

Today's passing roster

The measured table

Click any model for its endpoint and ready-to-run snippets in curl, Python, Go, Rust, and more.

latencymodelvendortool callfree nowagenticyesterday3 days agoa week ago
8.5aisingapore/Llama-SEA-LION-v3-70B-ITSEA-LIONuntestedDOWN—···
1.9north-mini-codeTokenReplyyesFREE3/5···
0.4mistral-code-latestMistral La PlateformeuntestedDOWN—···
0.5Qwen3-235B-A22BSambaNovauntestedDOWN—···
0.3nvidia/riva-translate-4b-instruct-v1.1nvidiauntestedDOWN—···
0.4meganova-ai/manta-flash-1.0MegaNova AIuntestedDOWN—✗··
0.3codestral-latestMistral La Plateformeyes (past)DOWN—···
3.9deepseek-ai/deepseek-v4-flash-0731NVIDIA NIMyesFREE3/5·✓✓
1.7openaiPollinations.aiyesFREE3/5···
0.3DeepSeek-V3.1sambanovauntestedCARD—✗✗✗
1.5qwen3.8-27bYolo-AutoyesFREE3/5···
0.201-ai/yi-largenvidiauntestedDOWN—✗✗✗
2.3cohere/north-mini-code:freeopenrouteryesFREE3/5✓✓✓
0.3DeepSeek-V3.1sambanovauntestedCARD—✗✗✗
29.6nvidia/nemotron-3-super-120b-a12bNVIDIA NIMyesFREE3/5···
0.3Meta-Llama-3.3-70B-InstructsambanovauntestedCARD—✗✗✗
0.2adept/fuyu-8bnvidiauntestedDOWN—✗✗✗
1.6dots-studio/dots-3-note-preview:freeopenrouteryesFREE3/5✓✓✓
0.3DeepSeek-V3.2sambanovauntestedCARD—✗✗✗
0.1ai21labs/jamba-1.5-large-instructnvidiauntestedDOWN—✗✗✗
0.2google/gemma-4-26b-a4b-it:freeopenrouteryes (past)DOWN—✗✗✗
0.3Meta-Llama-3.3-70B-InstructsambanovauntestedCARD—✗✗✗
0.3gpt-oss-120bSambaNovauntestedCARD—·✗✗
0.1aisingapore/sea-lion-7b-instructnvidiauntestedDOWN—✗✗✗
0.3google/gemma-4-31b-it:freeopenrouteryes (past)DOWN—✗✗✗
0.3MiniMax-M2.7sambanovauntestedCARD—✗✗✗
0.1bigcode/starcoder2-15bnvidiauntestedDOWN—✗✗✗
0.8inclusionai/ling-3.0-flash-fin:freeopenrouteryesFREE3/5✓✓✓
0.3MiniMax-M3sambanovauntestedCARD—✗✗✗

8 made a tool call, 0 answered in text only, of 29 discovered. A fail only means the endpoint wasn't free to use right then — it demanded a payment method, or it was down or erroring — so tool-call ability stayed untested, not disproven; a model that passed in a past report keeps its yes. Numbers are single-run measurements from one location — your mileage will vary. Prose by fallback (no model produced usable prose); the table is measured, never model-written. Reproduce with ./free-scan.sh from the engine repo.