Free LLM APIs (2026): free models, free credits & deals
As of 2026-10-17, 29 free models, credits and deals are listed: 29 active from official sources, 0 community reports (unverified).
Updated daily: free LLM API models, free tiers, signup credits and limited-time discounts from Gemini, Qwen, Kimi, GLM, OpenRouter and more — each with a source.
Free access by model: DeepSeek · GLM (Zhipu) · Step (StepFun) · Hunyuan / HY (Tencent) · MiMo (Xiaomi) · Claude · GPT (OpenAI) · Kimi (Moonshot) · Gemini · Qwen ·
Popular question: Is the DeepSeek API free? · Signup credits roundup
Source types: officialX postcommunity · unverified · Submit a deal →
Free models (20)
| Provider | Details | Dates | Status | Source |
|---|---|---|---|---|
| OpenRouter | Free on OpenRouter: NVIDIA Nemotron 3 Ultra (free) (context 1,000,000), rate-limited; #7 on OpenRouter weekly ranking
| — → — | active | official openrouter.ai · 2026-10-17 |
| Zhipu GLM (Z.ai) | GLM-4.7-Flash (200K ctx) and GLM-4.6V-Flash are free (input, output, cache)
| — → — | active | official open.bigmodel.cn · 2026-10-17 |
| Zhipu GLM (Z.ai) | OpenCode (official account): GLM 5.3 Flash is now free in OpenCode
Sign in to the OpenCode console, get an API key and call the model id; whether free models need credits first is not stated. | 2026-10-16 → — | active | official opencode.ai · 2026-10-17 |
| — | OpenCode (official account): Beam (1M ctx) free for a limited time in OpenCode
Sign in to the OpenCode console, get an API key and call the model id; whether free models need credits first is not stated. | 2026-10-16 → — | active | official opencode.ai · 2026-10-17 |
| DeepSeek | OpenCode Zen offers DeepSeek V4.1 Flash Free (model ID deepseek-v4.1-flash-free) free for a limited time — input/output/cache free; provider is zero-retention per the docs
Sign in to the OpenCode console, get an API key and call the model id; whether free models need credits first is not stated. | 2026-10-16 → — | active | official opencode.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: XingChen: Xing4.0 29B A4B (free) (context 262,144), rate-limited
| 2026-10-16 → — | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: NVIDIA: Nemotron 3.5 Super VL (free) (context 1,000,000), rate-limited
| 2026-10-15 → — | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: Dots Studio: Dots3-Note Preview (free) (context 512,000), rate-limited
| 2026-08-14 → 2026-12-31 | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: LiquidAI: LFM2.5-2.6B (free) (context 65,536), rate-limited
| 2026-08-11 → — | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: NVIDIA: Nemotron 3.5 Lightning (free) (context 1,000,000), rate-limited
| 2026-08-11 → — | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: Thinking Machines: Inkling Small (free) (context 1,048,576), rate-limited
| 2026-07-30 → — | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: Poolside: Laguna S 2.1 (free) (context 262,144), rate-limited
| 2026-07-21 → 2026-10-31 | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: Thinking Machines: Inkling (free) (context 1,048,576), rate-limited
| 2026-07-17 → — | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: Poolside: Laguna XS 2.1 (free) (context 262,144), rate-limited
| 2026-07-02 → 2026-10-31 | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: Cohere: North Mini Code (free) (context 256,000), rate-limited
| 2026-06-17 → — | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: NVIDIA: Nemotron 3.5 Content Safety (free) (context 128,000), rate-limited
| 2026-06-04 → — | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: NVIDIA: Nemotron 3 Nano Omni (free) (context 256,000), rate-limited
| 2026-04-28 → — | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: Google: Gemma 4 26B A4B (free) (context 262,144), rate-limited
| 2026-04-03 → — | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: Google: Gemma 4 31B (free) (context 262,144), rate-limited
| 2026-04-02 → — | active | official openrouter.ai · 2026-10-17 |
| OpenRouter | Free on OpenRouter: NVIDIA: Nemotron 3 Super (free) (context 262,144), rate-limited
| 2026-03-11 → — | active | official openrouter.ai · 2026-10-17 |
Free tiers (1)
| Provider | Details | Dates | Status | Source |
|---|---|---|---|---|
| Google Gemini | Gemini API free tier covers 3.8 Flash, 3.6 Flash, 3.5 Flash-Lite, 2.5 family (rate-limited; data may be used to improve Google products; 4 Argon excluded)
Free-tier data is used to improve Google products. | — → — | active | official ai.google.dev · 2026-10-17 |
Signup credits (4)
| Provider | Details | Dates | Status | Source |
|---|---|---|---|---|
| Tencent Hunyuan | First activation of Hunyuan text: 1M free tokens (shared across hunyuan-a13b etc.), valid 1 year
Official doc: real-name verification required; granted on first “Use now”. Hunyuan is migrating to TokenHub and the old platform “stops new purchases”, so whether new users can still claim is unverified. | — → — | active | official cloud.tencent.com · 2026-10-17 |
| ByteDance Doubao (Volcengine Ark) | 500K free inference tokens per Doubao model
Pricing page only says each model gets 500K free tokens; expiry and conditions are not stated. | — → — | active | official www.volcengine.com · 2026-10-17 |
| Alibaba Qwen (Model Studio) | 1M free tokens per model (qwen3.8-max, qwen3.8-flash, qwen3.7-plus…), 90 days, Beijing region only
Auto-granted when you first open Bailian; no real-name or purchase needed; unverified accounts stop when the quota runs out. Re-registering under the same identity cannot claim again. | — → — | active | official help.aliyun.com · 2026-10-17 |
| Moonshot AI (Kimi) | ¥15 voucher after identity verification (not usable for Kimi K3)
Granted after personal or company verification; expiry not stated. | — → — | active | official platform.kimi.com · 2026-10-17 |
Discounts (4)
| Provider | Details | Dates | Status | Source |
|---|---|---|---|---|
| Google Gemini | Gemini 4 Argon intro price $2/$10 until 2026-11-30, then $4/$20
| — → 2026-11-30 | active | official ai.google.dev · 2026-10-17 |
| Google Gemini | Gemini 3.8 Flash $0.75/$3.75 until 2026-12-31, doubles from 2027-01-01
| — → 2026-12-31 | active | official ai.google.dev · 2026-10-17 |
| Alibaba Qwen (Model Studio) | qwen3.7-plus shown as limited-time 20% off (end date not stated)
Paying requires real-name verification; Alipay top-up supported (Alibaba Cloud help). | — → — | active | official help.aliyun.com · 2026-10-17 |
| MiniMax | MiniMax-M3 permanently 50% off list (≤512K input: ¥2.10/¥8.40)
Balance top-up via WeChat Pay only; companies can wire (official FAQ). | — → — | active | official platform.minimaxi.com · 2026-10-17 |
Submit a deal
Know a free model, credit or discount? Send the official link or X post to deals@apiprice.app; we list it after verification.
FAQ
Which LLM APIs are free?
Zhipu GLM-4.7-Flash / GLM-4.6V-Flash are free; Gemini API has a free tier; OpenRouter lists several :free models; Alibaba, Doubao, Hunyuan and Kimi give signup credits.
Is the DeepSeek API free?
DeepSeek’s official API is pay-as-you-go. OpenCode Zen’s docs list DeepSeek V4.1 Flash Free (deepseek-v4.1-flash-free) as free for a limited time.