Free LLM APIs
As of Nov 22, 2026, 75 models compared on Sovyron have a free-tier API offer; Gemma 4 31B IT is free on the most providers (8).
The table lists the 20 models free on the most providers. Models with at least one free-tier (zero-price, non-subscription) offer. Free tiers often come with rate limits the catalog does not record.
| # | Model | Free on | Cheapest paid output |
|---|---|---|---|
| 1 | Gemma 4 31B IT | AIHubMix, AMD, Bothub, Google, Nvidia +3 more | $0.25 |
| 2 | GLM-4.7-Flash | AIHubMix, EmpirioLabs AI, Hugging Face, Kenari, Pendra +3 more | $0.4 |
| 3 | Nemotron 3 Ultra 550B A55B | AIHubMix, InferAll, Kenari, Nvidia, Requesty +1 more | $0.1 |
| 4 | GLM-5.3-Flash | AIHubMix, AMD, InferAll, Nvidia, OrcaRouter | $0.0185 |
| 5 | DeepSeek V4 Flash | InferX, Obscurify.ai, OrcaRouter, Pendra, SenseNova (China) | $0.132 |
| 6 | Nemotron 3.5 Lightning 30B A3B | AIHubMix, InferAll, Merge Gateway, Nvidia, Requesty | $0.096 |
| 7 | GLM-4.6V-Flash | EmpirioLabs AI, Kenari, Z.AI, ZenMux, Zhipu AI | $0.9 |
| 8 | DeepSeek V4.1 Flash | AMD, Nvidia, OpenCode, SenseNova (China) | $0.12 |
| 9 | GPT OSS 20B | InferAll, LMStudio, Nvidia, QVAC | $0.075 |
| 10 | Nemotron 3.5 Lightning | Kilo Gateway, OpenCode, OpenRouter, Vercel AI Gateway | $0.2 |
| 11 | GLM-5.3 | AIHubMix, InferAll, TokenRouter | $0.7013 |
| 12 | Kimi K3 | Kimi For Coding (kimi.com), Kimi For Coding (kimi.ai), SenseNova (China) | $4.5 |
| 13 | Qwen3.8 27B | AMD, Hetzner, IteraCompute | $0.32 |
| 14 | Gemma 4 26B A4B IT | AIHubMix, Google, NaN | $0.19 |
| 15 | Hy4 flash | Kenari, OpenCode, OpenCode Go | $0.1405 |
| 16 | Muse Glimmer 30B | Kenari, Nvidia, Requesty | $0.32 |
| 17 | Nemotron 3.5 Super VL | Kilo Gateway, OpenCode, OpenRouter | $0.35 |
| 18 | Nemotron 3 Ultra | Bothub, Kilo Gateway, OpenRouter | $1.7 |
| 19 | LongCat 2.5 | Kenari, OpenCode, OpenCode Go | $1.2 |
| 20 | Nemotron 3 Nano Omni | Kilo Gateway, Nvidia, OpenRouter | $0.9 |
FAQ
Which LLM APIs have a free tier?
As of Nov 22, 2026, 75 models compared on Sovyron have a free-tier API offer; Gemma 4 31B IT is free on the most providers (8).
Are free LLM API tiers really free?
They are listed at $0 per token by the provider. Providers may cap them with rate limits or daily quotas that the price catalog does not record, so check the provider before relying on one.
How is this list ranked?
By the number of providers that list a free tier for the model, then by how many providers sell it at a metered price.