TOOLS / FREE LLM MODEL

Free LLM Model

Browse documented API channels, check the free conditions, and follow a model-specific setup guide. No account connection or key upload required.

28 models · 28 documented channels Catalog reviewed 2026-09-29

Platform	Model	Source score	Context & capability	Free quota	Free period	Guide
NVIDIA NIM	DeepSeek V4.1 FlashDeepSeek	Unscored	1,000KCoding · Reasoning · Vision · Tool calling	Up to 40 requests/minute	No announced end date · subject to change	Use this model →
OpenRouter	inclusionAI: Ling 3.0 Flash SanteInclusionAI	95 ↗	262KReasoning · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
NVIDIA NIM	GLM 5.3 FlashZ.ai	Unscored	1,000KReasoning · Vision · Tool calling	Up to 40 requests/minute	No announced end date · subject to change	Use this model →
NVIDIA NIM	GLM 5.3Z.ai	Unscored	1,000KReasoning · Tool calling	Up to 40 requests/minute	No announced end date · subject to change	Use this model →
OpenRouter	Dots Studio: Dots3-Note PreviewDots Studio	69 ↗	512KReasoning · Vision · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
OpenRouter	Qwen: Qwen3.8 27BAlibaba / Qwen	93 ↗	262KCoding · Reasoning · Vision · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
OpenRouter	LiquidAI: LFM2.5-2.6BLiquid AI	62 ↗	66KReasoning · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
NVIDIA NIM	Nemotron 3.5 Lightning 30BNVIDIA	Unscored	1,000KCoding · Reasoning · Tool calling	Up to 40 requests/minute	No announced end date · subject to change	Use this model →
OpenRouter	NVIDIA: Nemotron 3.5 LightningNVIDIA	69 ↗	1,000KReasoning · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
NVIDIA NIM	Muse Glimmer 30BMeta	Unscored	131KReasoning · Vision · Tool calling	Up to 40 requests/minute	No announced end date · subject to change	Use this model →
OpenRouter	Thinking Machines: Inkling SmallThinking Machines Lab	82 ↗	1,049KCoding · Reasoning · Vision · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
OpenRouter	Poolside: Laguna S 2.1Poolside	67 ↗	262KCoding · Reasoning · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
OpenRouter	Thinking Machines: InklingThinking Machines Lab	80 ↗	1,049KCoding · Reasoning · Vision · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
NVIDIA NIM	Kimi K3Moonshot AI	Unscored	1,000KCoding · Reasoning · Vision · Tool calling	Up to 40 requests/minute	No announced end date · subject to change	Use this model →
NVIDIA NIM	Laguna XS 2.1Poolside	Unscored	262KCoding · Reasoning · Tool calling	Up to 40 requests/minute	No announced end date · subject to change	Use this model →
OpenRouter	Poolside: Laguna XS 2.1Poolside	66 ↗	262KCoding · Reasoning · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
OpenRouter	Cohere: North Mini CodeCohere	64 ↗	256KCoding · Reasoning · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
NVIDIA NIM	Nemotron 3 Ultra 550BNVIDIA	Unscored	1,000KCoding · Reasoning · Tool calling	Up to 40 requests/minute	No announced end date · subject to change	Use this model →
OpenRouter	NVIDIA: Nemotron 3 UltraNVIDIA	75 ↗	1,000KReasoning · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
Ollama Cloud	MiniMax M3MiniMax	Unscored	512KReasoning · Vision · Tool calling	Starter eligibility for this model unconfirmed; paid token rates are published	No announced end date · subject to change	Use this model →
OpenRouter	NVIDIA: Nemotron 3 Nano OmniNVIDIA	63 ↗	256KReasoning · Vision · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
Ollama Cloud	DeepSeek V4 ProDeepSeek	Unscored	1,000KReasoning · Tool calling	Starter eligibility for this model unconfirmed; paid token rates are published	No announced end date · subject to change	Use this model →
OpenRouter	Google: Gemma 4 26B A4B Google	63 ↗	262KReasoning · Vision · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
NVIDIA NIM	Gemma 4 31B ITGoogle	Unscored	262KCoding · Reasoning · Vision	Up to 40 requests/minute	No announced end date · subject to change	Use this model →
OpenRouter	Google: Gemma 4 31BGoogle	63 ↗	262KReasoning · Vision · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
NVIDIA NIM	Nemotron 3 Super 120BNVIDIA	Unscored	1,000KCoding · Reasoning · Tool calling	Up to 40 requests/minute	No announced end date · subject to change	Use this model →
OpenRouter	NVIDIA: Nemotron 3 SuperNVIDIA	59 ↗	262KReasoning · Tool calling	50 requests per day on the Free plan	No announced end date · subject to change	Use this model →
NVIDIA NIM	GPT OSS 20BOpenAI	Unscored	131KReasoning · Tool calling	Up to 40 requests/minute	No announced end date · subject to change	Use this model →

About this directory

OpenRouter IDs and zero prompt/completion prices came from its public model API. NVIDIA and Ollama Cloud rows use provider model IDs and endpoints checked against their official pages, with free-tier listings cross-checked against freeLLM. Ollama's official Free plan includes starter usage for a smaller set of models, so eligibility for a particular model must be checked in the signed-in account. “Docs checked” means source review, not a successful model response. A missing end date is not a promise of permanent access.

Coverage in this first pass: 15 OpenRouter channels, 11 NVIDIA NIM channels, and 2 Ollama Cloud channels. The “No phone” condition is optional because NVIDIA's current provider snapshot marks phone verification as required. Turn it on when that requirement matters to you.

Start with Qwen3.8 27B, or filter Platform to NVIDIA NIM or Ollama Cloud to open the new guides.