anytimelime — the free AI field report

2026-09-28 · generated by a free model interviewing free models · anytimelime, formerly Project42 · get the engine →

What's free in AI right now — 2026-09-28

Debut. Two models finally answered tool calls today, breaking a long silence. meta-llama/Llama-3.3-70B-Instruct and meta/llama-3.2-90b-vision-instruct both logged a tool response for the first time ever, joining the tool‑call club after months of being listed as free but never responding. Their arrivals are a quiet triumph for anyone who kept the roster alive.

Running. The free tier feels like a staggered lineup. A handful of standouts dominate the latency charts:

meta-llama/Llama-3.3-70B-Instruct – 0.3 s, tool‑yes (the speediest of the new debuts)
meta/llama-3.2-90b-vision-instruct – 44.9 s, tool‑yes (a heavyweight vision model finally usable)
aion-labs/aion-3.0 – 1.4 s, tool‑yes (steady workhorse)
openai-fast – sub‑second, tool‑yes (the classic quick pick for low‑cost tasks)

Other free fixtures continue to hum under a second: deepseek-ai/DeepSeek-V3-0324, cohere/north-mini-code:free, and qwen/qwen3.5-4b-free. The latency spread—from fractions of a second to almost a minute—means any production script must favor speed first and then roll forward before the next lock.

Meaning. The churn is the headline. SambaNova’s free lane suffered a quiet but sweeping lock this morning: half‑dozen models that answered tools just days ago (DeepSeek-V3.1, MiniMax-M3, DeepSeek-V3.2, gemma-4-31B-it, MiniMax-M2.7, Meta-Llama-3.3-70B-Instruct, gpt-oss-120b) now sit behind a “payment method required” screen, even though the probe never reached the model itself. Meanwhile, the roster shuffles: poolside/laguna-s-2.1:free and nvidia/nemotron-3-super-120b-a12b:free have vanished from the tool‑call club, while nvidia/ising-calibration-1.5-31b and nvidia/nemotron-3-nano-omni-30b-a3b-reasoning keep humming. For builders on free tiers the lesson is simple—rotation isn’t a convenience but a survival tactic; a model that answered yesterday can be silent tomorrow, and a “free” label can be stripped without warning.

Running it yourself. Clone the repo from github.com/fusionpoweredsoftware/anytimelime, set LIME_KEY to your OpenRouter key, run ./lime.sh. It pulls 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.

Today's passing roster

openai — Pollinations · 0.3s · agentic 3/5
openai — Pollinations · 0.3s · agentic 3/5
meta-llama/Llama-3.3-70B-Instruct — Hugging Face Inference · 0.7s · agentic 3/5
deepseek-ai/DeepSeek-V3-0324 — HuggingFace Inference Router · 0.7s · agentic 3/5
liquid/lfm-2.5-2.6b:free — openrouter · 0.7s · agentic 3/5
meta/llama-3.2-11b-vision-instruct — nvidia · 0.7s · agentic 3/5
nvidia/ising-calibration-1.5-31b — nvidia · 0.7s · agentic 3/5
nvidia/nemotron-3-super-120b-a12b — nvidia · 0.7s · agentic 3/5
openai/gpt-oss-120b:free — FastRouter · 0.8s · agentic 3/5
typhoon-v2.5-30b-a3b-instruct — Typhoon (SCB 10X) · 0.8s · agentic 3/5
inclusionai/ling-3.0-flash-fin:free — openrouter · 0.8s · agentic 3/5
inclusionai/ling-3.0-flash-sante:free — openrouter · 0.9s · agentic 3/5
google/diffusiongemma-26b-a4b-it — nvidia · 0.9s · agentic 3/5
cohere/north-mini-code:free — openrouter · 1.1s · agentic 3/5
qwen3-8b — Free.ai · 1.1s · agentic 3/5
openrouter/free — openrouter · 1.2s · agentic 3/5
openai/gpt-oss-20b:free — FastRouter · 1.3s · agentic 3/5
nvidia/nemotron-3-ultra-550b-a55b:free — openrouter · 1.3s · agentic 3/5
Qwen/Qwen3-Coder-480B-A35B-Instruct — HuggingFace Router · 1.4s · agentic 3/5
aion-labs/aion-3.0 — AION Labs · 1.4s · agentic 3/5
openai-fast — Pollinations · 1.4s · agentic 3/5
google/gemma-4-31b-it — nvidia · 1.4s · agentic 3/5
stealth/space-bunny-alpha — openrouter · 1.4s · agentic 3/5
nvidia/nemotron-3-ultra-550b-a55b — nvidia · 1.5s · agentic 3/5
qwen/qwen3.5-4b-free — Mixlayer · 1.7s · agentic 3/5
dots-studio/dots-3-note-preview:free — openrouter · 1.7s · agentic 3/5
north-mini-code — TokenReply · 1.9s · agentic 3/5
zai-org/GLM-5.3-Flash — Gonka Broker · 2.4s · agentic 3/5
nemotron-3-super-120b-a12b:free — Naga AI · 2.5s · agentic 3/5
nvidia/nemotron-3.5-lightning:free — openrouter · 2.5s · agentic 3/5
MiniMax-M2.7-Official — Pooled AI · 2.8s · agentic 3/5
nvidia/nemotron-3-nano-omni-30b-a3b-reasoning — nvidia · 3.0s · agentic 3/5
nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free — openrouter · 3.1s · agentic 3/5
meta/muse-glimmer-30b — nvidia · 3.1s · agentic 3/5
intern-s2-preview — Intern AI · 3.6s · agentic 3/5
MiniMaxAI/MiniMax-M2.7 — Gonka Broker · 7.6s · agentic 3/5
agnes-3.0-flash — Agnes AI · 9.9s · agentic 3/5
nvidia/nemotron-3.5-lightning-30b-a3b — nvidia · 25.5s · agentic 3/5
meta/llama-3.2-90b-vision-instruct — nvidia · 44.9s · agentic 3/5
glm-4.7-flash — Zhipu AI (Z.ai) · 47.3s · agentic 3/5
glm-4.7-flash — Zhipu AI (Z.ai) · 47.3s · agentic 3/5

The measured table

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

latency	model	vendor	tool call	free now	agentic	yesterday	3 days ago	a week ago
—	`@cf/google/gemma-3-12b-it`	PersorAI	untested	DOWN	—	·	·	·
—	`@cf/google/gemma-3-12b-it`	PersorAI	untested	DOWN	—	·	·	·
0.3	`DeepSeek-V4-Flash`	DAHL (Gonka Network, promo)	untested	DOWN	—	·	·	·
—	`Mistral-7B-Instruct-v0.3`	OVHcloud AI Endpoints	untested	DOWN	—	·	·	·
—	`Qwen/Qwen3.6-35B-A3B-FP8`	Hetzner AI Inference	untested	DOWN	—	·	·	·
0.2	`deepseek-ai/deepseek-v3.1`	NVIDIA NIM	untested	DOWN	—	·	·	·
0.7	`meta-llama/Llama-3.3-70B-Instruct`	Hugging Face Inference	yes	FREE	3/5	·	·	·
0.3	`open-mistral-nemo`	Mistral AI	untested	DOWN	—	·	·	·
0.5	`@cf/mistral/mistral-small-3.1-24b-instruct`	PersorAI	untested	DOWN	—	·	·	·
0.4	`Llama-4-Maverick`	SambaNova	untested	DOWN	—	·	·	·
—	`@cf/qwen/qwen3-30b-a3b-fp8`	PersorAI	untested	DOWN	—	·	·	·
0.4	`Llama-4-Scout`	SambaNova	untested	DOWN	—	·	·	·
—	`@cf/qwen/qwen3-30b-a3b-fp8`	PersorAI	untested	DOWN	—	·	·	·
0.5	`openai-chat`	PersorAI	untested	DOWN	—	·	·	·
0.2	`01-ai/yi-large`	nvidia	untested	DOWN	—	✗	✗	✗
0.3	`DeepSeek-V3.1`	SambaNova	untested	CARD	—	✗	✗	✗
2.8	`MiniMax-M2.7-Official`	Pooled AI	yes	FREE	3/5	✓	✓	·
0.3	`MiniMax-M3`	sambanova	untested	CARD	—	✗	✗	✗
7.6	`MiniMaxAI/MiniMax-M2.7`	Gonka Broker	yes	FREE	3/5	✓	✓	✓
1.4	`Qwen/Qwen3-Coder-480B-A35B-Instruct`	HuggingFace Router	yes	FREE	3/5	✓	·	·
9.9	`agnes-3.0-flash`	Agnes AI	yes	FREE	3/5	✓	✓	·
1.4	`aion-labs/aion-3.0`	AION Labs	yes	FREE	3/5	✓	✓	·
2.3	`aisingapore/Qwen-SEA-LION-v4.5-27B-IT`	SEA-LION (AI Singapore)	yes (past)	DOWN	—	✗	✗	✗
0.3	`codestral-latest`	Mistral	yes (past)	DOWN	—	✗	✗	✗
1.1	`cohere/north-mini-code:free`	openrouter	yes	FREE	3/5	✓	✓	✓
0.7	`deepseek-ai/DeepSeek-V3-0324`	HuggingFace Inference Router	yes	FREE	3/5	✓	✓	·
0.5	`deepseek-ai/DeepSeek-V3.1`	SiliconFlow	yes (past)	DOWN	—	✗	✗	·
1.7	`deepseek-ai/DeepSeek-V3`	SiliconFlow	untested	DOWN	—	✗	·	·
0.3	`gemini-2.5-flash`	Google Gemini (OpenAI-compat)	untested	DOWN	—	·	·	·
0.6	`glm-4.5-flash`	Zhipu AI (Z.ai)	yes (past)	DOWN	—	✗	✗	·
47.3	`glm-4.7-flash`	Zhipu AI (Z.ai)	yes	FREE	3/5	✗	✗	·
0.9	`intern-latest`	Intern AI	yes (past)	DOWN	—	✗	✗	·
0.8	`minimax/minimax-m3:free`	xKiro AI	yes (past)	DOWN	—	✗	✗	·
0.3	`mistral-small-latest`	Mistral AI	untested	DOWN	—	✗	·	·
2.5	`nemotron-3-super-120b-a12b:free`	Naga AI	yes	FREE	3/5	✓	✓	·
1.9	`north-mini-code`	TokenReply	yes	FREE	3/5	✓	✓	✓
1.4	`openai-fast`	Pollinations	yes	FREE	3/5	·	✓	✓
0.8	`openai/gpt-oss-120b:free`	FastRouter	yes	FREE	3/5	✓	✓	·
0.3	`openai`	Pollinations	yes	FREE	3/5	✓	✓	✓
1.7	`qwen/qwen3.5-4b-free`	Mixlayer	yes	FREE	3/5	✓	✓	·
1.1	`qwen3-8b`	Free.ai	yes	FREE	3/5	✓	✓	·
0.3	`qwen3.8-27b`	Yolo-Auto	yes (past)	DOWN	—	✗	✗	✗
60.1	`tokeness/free`	Tokeness	yes (past)	DOWN	—	✗	✗	✗
0.8	`typhoon-v2.5-30b-a3b-instruct`	Typhoon (SCB 10X)	yes	FREE	3/5	✓	✓	·
0.2	`adept/fuyu-8b`	nvidia	untested	DOWN	—	✗	✗	✗
—	`@cf/meta/llama-3.3-70b-instruct-fp8-fast`	PersorAI	untested	DOWN	—	·	·	·
—	`@cf/meta/llama-3.3-70b-instruct-fp8-fast`	PersorAI	untested	DOWN	—	·	·	·
0.3	`DeepSeek-V3.2`	SambaNova	untested	CARD	—	✗	✗	✗
0.5	`GLM-5.3-Flash`	DAHL (Gonka Network, promo)	untested	DOWN	—	·	·	·
0.3	`gemma-4-31B-it`	sambanova	untested	CARD	—	✗	✗	✗
0.5	`deepseek-ai/DeepSeek-V4-Flash-0731`	Gonka Broker	yes (past)	DOWN	—	✗	✗	✗
1.7	`dots-studio/dots-3-note-preview:free`	openrouter	yes	FREE	3/5	✓	✓	✓
1.7	`deepseek-ai/DeepSeek-V3`	SiliconFlow	untested	DOWN	—	✗	·	·
0.1	`deepseek-ai/deepseek-v4-flash-0731`	NVIDIA NIM	yes (past)	DOWN	—	✗	✗	✗
47.3	`glm-4.7-flash`	Zhipu AI (Z.ai)	yes	FREE	3/5	✗	✗	·
3.6	`intern-s2-preview`	Intern AI	yes	FREE	3/5	✓	·	·
0.3	`openai`	Pollinations	yes	FREE	3/5	✓	✓	✓
1.3	`openai/gpt-oss-20b:free`	FastRouter	yes	FREE	3/5	✓	·	·
0.2	`ai21labs/jamba-1.5-large-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
—	`@cf/openai/gpt-oss-120b`	PersorAI	untested	DOWN	—	·	·	·
0.3	`MiniMax-M2.7`	SambaNova	untested	CARD	—	✗	✗	✗
2.4	`zai-org/GLM-5.3-Flash`	Gonka Broker	yes	FREE	3/5	✓	✗	✓
0.2	`google/gemma-4-26b-a4b-it:free`	openrouter	yes (past)	DOWN	—	✗	✗	·
0.3	`inclusionAI/Ling-flash-2.0`	SiliconFlow	yes (past)	DOWN	—	✗	✗	·
0.1	`meta/llama-3.3-70b-instruct`	NVIDIA NIM	untested	DOWN	—	✗	·	·
0.2	`aisingapore/sea-lion-7b-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
—	`@cf/openai/gpt-oss-120b`	PersorAI	untested	DOWN	—	·	·	·
0.4	`google/gemma-4-31b-it:free`	openrouter	yes (past)	DOWN	—	✗	✗	·
0.2	`nvidia/llama-3.1-nemotron-70b-instruct`	NVIDIA NIM	untested	DOWN	—	✗	✗	✗
0.1	`bigcode/starcoder2-15b`	nvidia	untested	DOWN	—	✗	✗	✗
0.3	`Meta-Llama-3.3-70B-Instruct`	SambaNova	untested	CARD	—	✗	✗	✗
0.8	`inclusionai/ling-3.0-flash-fin:free`	openrouter	yes	FREE	3/5	✓	✓	✓
0.2	`databricks/dbrx-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
0.5	`@cf/zai-org/glm-5.2`	PersorAI	untested	DOWN	—	✗	·	·
0.3	`MiniMax-M2.7`	SambaNova	untested	CARD	—	✗	✗	✗
0.9	`inclusionai/ling-3.0-flash-sante:free`	openrouter	yes	FREE	3/5	✓	✓	✓
0.1	`deepseek-ai/deepseek-coder-6.7b-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
0.3	`Qwen3-235B`	SambaNova	untested	DOWN	—	·	·	·
0.7	`liquid/lfm-2.5-2.6b:free`	openrouter	yes	FREE	3/5	✗	✓	✓
60.1	`deepseek-ai/deepseek-v4.1-flash`	nvidia	untested	DOWN	—	✗	✗	·
0.3	`gemma-3-12b-it`	SambaNova	untested	DOWN	—	·	·	·
3.1	`nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free`	openrouter	yes	FREE	3/5	✓	✓	·
0.1	`google/codegemma-1.1-7b`	nvidia	untested	DOWN	—	✗	✗	✗
0.3	`gpt-oss-120b`	SambaNova	untested	CARD	—	✗	✗	✗
0.5	`nvidia/nemotron-3-super-120b-a12b:free`	openrouter	yes (past)	DOWN	—	✓	✓	·
0.2	`google/codegemma-7b`	nvidia	untested	DOWN	—	✗	✗	✗
1.3	`nvidia/nemotron-3-ultra-550b-a55b:free`	openrouter	yes	FREE	3/5	✓	✓	✓
0.2	`google/deplot`	nvidia	untested	DOWN	—	✗	✗	✗
0.2	`nvidia/nemotron-3.5-content-safety:free`	openrouter	untested	DOWN	—	✗	✗	✗
0.9	`google/diffusiongemma-26b-a4b-it`	nvidia	yes	FREE	3/5	✓	✓	✓
2.5	`nvidia/nemotron-3.5-lightning:free`	openrouter	yes	FREE	3/5	✓	✓	✓
0.2	`google/gemma-2b`	nvidia	untested	DOWN	—	✗	✗	✗
1.2	`openrouter/free`	openrouter	yes	FREE	3/5	✓	✓	✓
0.1	`google/gemma-3-12b-it`	nvidia	untested	DOWN	—	✗	✗	✗
0.2	`poolside/laguna-s-2.1:free`	openrouter	yes (past)	DOWN	—	✓	✓	✓
0.1	`google/gemma-3-4b-it`	nvidia	untested	DOWN	—	✗	✗	✗
0.2	`poolside/laguna-xs-2.1:free`	openrouter	yes (past)	DOWN	—	✗	✗	✓
1.4	`google/gemma-4-31b-it`	nvidia	yes	FREE	3/5	✓	✓	✓
0.3	`qwen/qwen3.8-27b:free`	openrouter	yes (past)	DOWN	—	✗	✗	·
0.1	`google/recurrentgemma-2b`	nvidia	untested	DOWN	—	✗	✗	✗
1.4	`stealth/space-bunny-alpha`	openrouter	yes	FREE	3/5	✓	✓	·
0.1	`ibm/granite-3.0-3b-a800m-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`ibm/granite-3.0-8b-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`ibm/granite-34b-code-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`ibm/granite-8b-code-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`meta/codellama-70b`	nvidia	untested	DOWN	—	✗	✗	✗
0.7	`meta/llama-3.2-11b-vision-instruct`	nvidia	yes	FREE	3/5	✓	✓	✓
44.9	`meta/llama-3.2-90b-vision-instruct`	nvidia	yes	FREE	3/5	✗	✗	✗
0.2	`meta/llama2-70b`	nvidia	untested	DOWN	—	✗	✗	✗
3.1	`meta/muse-glimmer-30b`	nvidia	yes	FREE	3/5	✓	✓	✓
0.1	`microsoft/kosmos-2`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`microsoft/phi-3-vision-128k-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`microsoft/phi-3.5-moe-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`mistralai/codestral-22b-instruct-v0.1`	nvidia	untested	DOWN	—	✗	✗	✗
0.2	`mistralai/mistral-7b-instruct-v0.3`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`mistralai/mistral-large-2-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`mistralai/mistral-large`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`mistralai/mixtral-8x22b-v0.1`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`moonshotai/kimi-k2.6`	nvidia	untested	DOWN	—	✗	✗	✗
60.1	`moonshotai/kimi-k3`	nvidia	yes (past)	DOWN	—	✓	✓	✗
0.2	`nv-mistralai/mistral-nemo-12b-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
0.3	`nvidia/ai-synthetic-video-detector`	nvidia	untested	DOWN	—	✗	✗	✗
0.2	`nvidia/cosmos-reason2-8b`	nvidia	untested	DOWN	—	✗	✗	✗
0.7	`nvidia/ising-calibration-1.5-31b`	nvidia	yes	FREE	3/5	✓	✓	✗
0.2	`nvidia/llama-3.1-nemotron-51b-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
0.2	`nvidia/llama-3.1-nemotron-ultra-253b-v1`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`nvidia/llama3-chatqa-1.5-70b`	nvidia	untested	DOWN	—	✗	✗	✗
0.2	`nvidia/mistral-nemo-minitron-8b-8k-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
3.0	`nvidia/nemotron-3-nano-omni-30b-a3b-reasoning`	nvidia	yes	FREE	3/5	✗	✓	✗
0.7	`nvidia/nemotron-3-super-120b-a12b`	nvidia	yes	FREE	3/5	✗	✗	✗
1.5	`nvidia/nemotron-3-ultra-550b-a55b`	nvidia	yes	FREE	3/5	✓	✓	✗
0.2	`nvidia/nemotron-3.5-content-safety`	nvidia	untested	DOWN	—	✗	✗	✗
25.5	`nvidia/nemotron-3.5-lightning-30b-a3b`	nvidia	yes	FREE	3/5	✓	✓	✗
0.2	`nvidia/nemotron-4-340b-instruct`	nvidia	untested	DOWN	—	✗	✗	✗
0.2	`nvidia/nemotron-4-340b-reward`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`nvidia/nemotron-nano-3-30b-a3b`	nvidia	untested	DOWN	—	✗	✗	✗
0.2	`nvidia/nemotron-parse-2.0`	nvidia	untested	DOWN	—	✗	✗	✗
0.2	`nvidia/nemotron-parse`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`nvidia/neva-22b`	nvidia	untested	DOWN	—	✗	✗	✗
0.1	`nvidia/nvclip`	nvidia	untested	DOWN	—	✗	✗	✗
0.7	`nvidia/riva-translate-4b-instruct-v1.1`	nvidia	untested	DOWN	—	✗	✗	✗
60.1	`nvidia/riva-translate-4b-instruct-v2`	nvidia	untested	DOWN	—	·	·	✗
0.1	`nvidia/riva-translate-4b-instruct`	nvidia	untested	DOWN	—	✗	✗	✗

41 made a tool call, 0 answered in text only, of 143 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 cohere/north-mini-code:free; the table is measured, never model-written. Reproduce with ./free-scan.sh from the engine repo.