Free Ultra-budget open-model inference for AI agents

Run it free (12)

• Qwen3-0.6B-Base · open_weights · oss · Apache-2.0

  Downloadable Qwen3 0.6B base weights. This row is this checkpoint only, not every Qwen model.

  Limits: none stored.

  Source · asOf 2026-10-25 · LICENSE file fetched 2026-10-03 is Apache License 2.0. Weights are the model repo, not this license URL. LICENSE re-fetched twice 2026-10-25 (HTTP 200, byte-identical): 'Apache License Version 2.0, January 2004'.

• DeepSeek-V3.2 weights · open_weights · oss · MIT

  Downloadable DeepSeek-V3.2 weights from the Hugging Face repo whose LICENSE file is MIT.

  Limits: none stored.

  Source · asOf 2026-10-25 · This is DeepSeek-V3.2, not DeepSeek-V3. V3 used a separate LICENSE-MODEL and is not indexed here. LICENSE re-fetched twice 2026-10-25 (HTTP 200, byte-identical): 'MIT License'.

• Ollama · local_runtime · oss · MIT

  Local model runner and OpenAI-compatible server. You still supply weights and your own hardware.

  Limits: none stored.

  Source · asOf 2026-10-25 · Re-fetched twice 2026-10-25 (~10 s apart, plain curl, no login): the GitHub blob page answered HTTP 200 both times (page chrome differs between fetches), so the raw file https://raw.githubusercontent.com/ollama/ollama/main/LICENSE was also fetched twice (HTTP 200, byte-identical): 'MIT License' / 'Copyright (c) Ollama'. License unchanged.

• llama.cpp · local_runtime · oss · MIT

  C/C++ GGUF inference CLI and server. Weights are not included.

  Limits: none stored.

  Source · asOf 2026-10-25 · LICENSE fetched 2026-10-03, copyright 2023-2026 The ggml authors. Re-fetched twice 2026-10-25 (~10 s apart, plain curl, no login): the GitHub blob page answered HTTP 200 both times (page chrome differs between fetches), so the raw file https://raw.githubusercontent.com/ggml-org/llama.cpp/master/LICENSE was also fetched twice (HTTP 200, byte-identical): 'MIT License' / 'Copyright (c) 2023-2026 The ggml authors'. License unchanged.

• Cloudflare Workers AI daily neurons · free_inference · free

  10,000 Neurons per day at no charge on Workers Free and Workers Paid. Workers Free cannot buy more. The pool is shared across models.

  Limits: neuronsPerDay: 10000 · resets: 00:00 UTC · sharedAcrossModels: true

  Source · asOf 2026-10-25 · Fetched twice 2026-10-25 about 10 s apart without logging in (HTTP 200, no redirect, identical HTML; the page shows 'Last updated Oct 20, 2026'). 'Our free allocation allows anyone to use a total of 10,000 Neurons per day at no charge.' 'To use more than 10,000 Neurons per day, you need to sign up for the Workers Paid plan.' Free allocation table: 'Workers Free 10,000 Neurons per day N/A - Upgrade to Workers Paid' and 'Workers Paid 10,000 Neurons per day $0.011 / 1,000 Neurons'. 'All limits reset daily at 00:00 UTC.' Paid usage above the allocation is a price row, not a free cap. 'Some models require a paid billing method.' The page lists @cf/moonshotai/kimi-k2.6, @cf/moonshotai/kimi-k2.7-code, @cf/zai-org/glm-5.2, @cf/zai-org/glm-5.3, @cf/zai-org/glm-5.3-flash, @cf/deepseek-ai/deepseek-v4-flash-0731, @cf/deepseek-ai/deepseek-v4.1-flash-1017 and @cf/deepseek-ai/deepseek-v4-pro-0813 ('You can access these models with either the Workers Paid plan or prepaid AI Gateway credits'); these are not free. Per-model token prices are not a $0 SKU.

• OpenRouter :free variants and openrouter/free · free_inference · free

  Model ids ending in :free, and the openrouter/free router, are priced at zero on OpenRouter. Which models have a free entry changes.

  Limits: appliesTo: model variants with an ID ending in :free · requestsPerMinute: 20 · requestsPerDayCreditsPurchasedLessThan10: 50 · requestsPerDayCreditsPurchasedAtLeast10: 1000 · dailyWindow: UTC day

  Source · asOf 2026-10-25 · Fetched twice 2026-10-25 about 10 s apart without logging in: https://openrouter.ai/docs/api/reference/limits answers HTTP 308 to https://openrouter.ai/docs/api_reference/limits (HTTP 200; same tag-stripped text both times; no last-modified header; page metadata dateModified 2026-10-10T16:11:54Z). The HTML table cells are empty until scripts run (the page defines the values as constants), so the numbers are taken from the page's markdown view https://openrouter.ai/docs/api_reference/limits.md (HTTP 200, fetched twice, identical): 'Free usage limits: If you're using a free model variant (with an ID ending in `:free`), the following limits apply:' table 'Credits purchased (all time) | Requests per minute | Requests per day', 'Less than 10 | 20 | 50', 'At least 10 | 20 | 1000'. 'Free-model requests allowed per UTC day (see the free usage limits below)'. 'On free variants, purchase at least 10 credits to raise your daily limit, or switch to the paid variant of the model, which has no platform-level request cap.' 'If your account has a negative credit balance, you may see errors, including for free models.' (the HTML renders the status code as a component: 402 Payment Required). The one-credit-below-threshold grace is shown only as the unrendered placeholder '{FREE_MODEL_CREDITS_THRESHOLD - 1}', so it is not stored. The page does not say whether openrouter/free draws on the same caps. Third-party tables were not copied. Do not copy OpenRouter prices or limits onto other Depot providers.

• Qwen3-8B · open_weights · oss · Apache-2.0

  Downloadable Qwen3-8B weights.

  Limits: none stored.

  Source · asOf 2026-10-25 · LICENSE file fetched 2026-10-03 is Apache License 2.0. Not a claim about every Qwen3 checkpoint. LICENSE re-fetched twice 2026-10-25 (HTTP 200, byte-identical): 'Apache License Version 2.0, January 2004'.

• Phi-4 · open_weights · oss · MIT

  Downloadable Microsoft Phi-4 weights.

  Limits: none stored.

  Source · asOf 2026-10-25 · LICENSE file fetched 2026-10-03 is the MIT License. LICENSE re-fetched twice 2026-10-25 (HTTP 200, byte-identical): 'MIT License'.

• gpt-oss-20b · open_weights · oss · Apache-2.0

  Downloadable OpenAI gpt-oss-20b weights.

  Limits: none stored.

  Source · asOf 2026-10-25 · LICENSE file fetched 2026-10-03 is Apache License 2.0. LICENSE re-fetched twice 2026-10-25 (HTTP 200, byte-identical): 'Apache License Version 2.0, January 2004'.

• SmolLM2-1.7B-Instruct · open_weights · oss · Apache-2.0

  Downloadable SmolLM2 1.7B Instruct weights.

  Limits: none stored.

  Source · asOf 2026-10-25 · Model card front matter fetched 2026-10-03 says license: apache-2.0. A separate LICENSE file was not in the repo (404). README re-fetched twice 2026-10-25 (HTTP 200, byte-identical): front matter 'license: apache-2.0'; raw/main/LICENSE still 404 (twice).

• Mistral-7B-Instruct-v0.3 · open_weights · oss · Apache-2.0

  Downloadable Mistral-7B-Instruct-v0.3 weights.

  Limits: none stored.

  Source · asOf 2026-10-25 · Model card front matter fetched 2026-10-03 says license: apache-2.0. A separate LICENSE file was 404. README re-fetched twice 2026-10-25 (HTTP 200, byte-identical): front matter 'license: apache-2.0'; raw/main/LICENSE still 404 (twice).

• SGLang · local_runtime · oss · Apache-2.0

  Open-source serving runtime for LLMs.

  Limits: none stored.

  Source · asOf 2026-10-25 · LICENSE fetched 2026-10-03 is Apache License 2.0. Re-fetched twice 2026-10-25 (~10 s apart, plain curl, no login): HTTP 200, byte-identical: 'Apache License Version 2.0, January 2004'. License unchanged.

Reference (docs, public datasets and benchmarks) (0)

None.