# Free alternatives to Phi-4 Reasoning

Phi-4 Reasoning: Big reasoning power, small models

24 products that do a similar job and cost nothing to start: 20 free outright, 4 with a free plan. Closest in meaning first. Each price was read from the product's own site on the date shown.

Page: https://nowfound.app/free-alternatives/phi-4-reasoning
All alternatives, paid too: https://nowfound.app/alternatives/phi-4-reasoning.md
About Phi-4 Reasoning: https://nowfound.app/launch/phi-4-reasoning.md

| # | Product | Price | Checked | What it does |
|---|---|---|---|---|
| 1 | [Google Gemma 4](https://nowfound.app/launch/google-gemma-4) | Free | 2026-11-21 | Google's most intelligent open models to date |
| 2 | [DeepSeek R1](https://nowfound.app/launch/deepseek-r1) | Free | 2026-11-21 | Free, Fastest, 671B Full Version, No Sign-Up |
| 3 | [guppylm](https://nowfound.app/launch/i-built-a-tiny-llm-to-demystify-how-language-models-work) | Free | 2026-11-21 | Built a ~9M param LLM from scratch to understand how they actually work. Vanilla transformer, 60K synthetic conversations, ~130 lines of PyTorch. Trains in 5 m… |
| 4 | [Formal Verification for Machine Learning Models](https://nowfound.app/launch/formal-verification-for-machine-learning-models-using-lean-4) | Free | 2026-11-21 | Framework for specifying and proving properties—such as robustness, fairness, and interpretability—of machine learning models using Lean 4. |
| 5 | [Magistral](https://nowfound.app/launch/magistral) | Free plan, paid from $5.99/mo | 2026-11-21 | The first reasoning model by Mistral AI |
| 6 | [DeepSeek-VL2](https://nowfound.app/launch/deepseek-vl2) | Free | 2026-11-21 | MoE vision-language, now easier to access |
| 7 | [Hierarchical Reasoning Model](https://nowfound.app/launch/hierarchical-reasoning-model) | Free | 2026-11-21 | Brain-inspired, multi-level reasoning & planning AI model |
| 8 | [Open-source engine running Gemma 4 26B in 2 GB](https://nowfound.app/launch/open-source-engine-running-gemma-4-26b-in-2-gb-ram-on-any-m) | Free | 2026-11-20 | TurboFieldfare is an inference engine built in Swift and Metal that enables M-series Macs to run the Gemma 4 26B model using approximately 2 GB of RAM. |
| 9 | [NLP inside your database](https://nowfound.app/launch/nlp-inside-your-database) | Free plan, paid from $400/mo | 2026-11-22 | Query OpenAI’s GPT-3 from directly inside your database |
| 10 | [Z80-μLM](https://nowfound.app/launch/z80-lm-a-conversational-ai-that-fits-in-40kb) | Free | 2026-11-21 | Z80-μLM is a character-level language model compressed into a 40KB file that runs on a Z80 processor with 64KB RAM, executable in CP/M emulators or on vintage… |
| 11 | [Entropy-Guided Loop](https://nowfound.app/launch/entropy-guided-loop-how-to-make-small-models-reason) | Free | 2026-11-21 | TLDR: A small, vendor-agnostic inference loop that turns token logprobs/perplexity/entropy into an extra pass and reasoning for LLMs. - Captures logprobs/top-k… |
| 12 | [Duplicate 3 layers in a 24B LLM, logical](https://nowfound.app/launch/duplicate-3-layers-in-a-24b-llm-logical-deduction-22-76-no-t) | Free | 2026-11-21 | This project demonstrates a technique for improving large language model reasoning by duplicating specific layer blocks without retraining. |
| 13 | [GPT4All](https://nowfound.app/launch/gpt4all) | Free | 2026-11-21 | A chatbot trained on a massive collection of clean data |
| 14 | [MiniCPM 4.0](https://nowfound.app/launch/minicpm-4-0) | Free | 2026-11-21 | Ultra-efficient on-device AI, now even faster |
| 15 | [Agno](https://nowfound.app/launch/agno) | Free | 2026-11-21 | Build, run, and manage agent platforms. |
| 16 | [Zro](https://nowfound.app/launch/zro) | Free plan, paid from $48/mo | 2026-11-20 | Private inference for coding agents |
| 17 | [Auto-GPT](https://nowfound.app/launch/auto-gpt) | Free | 2026-11-21 | An Autonomous GPT-4 Experiment |
| 18 | [Evidently AI](https://nowfound.app/launch/evidently-ai) | Free | 2026-09-30 | Open-source evaluations and observability for LLM apps |
| 19 | [Moonshine Open-Weights STT models](https://nowfound.app/launch/moonshine-open-weights-stt-models-higher-accuracy-than-whisp) | Free | 2026-11-21 | Moonshine offers open-weights speech-to-text models that achieve lower word-error rates than OpenAI's Whisper Large v3, built by a six-person startup. |
| 20 | [Omnigraph](https://nowfound.app/launch/omnigraph) | Free | 2026-11-13 | Open source graph database for your AI agents |
| 21 | [Phidata](https://nowfound.app/launch/phidata-build-ai-agents-with-memory-knowledge-tools-and-reas) | Free | 2026-11-21 | Phidata is a framework for building AI agents that can retain memory, access knowledge bases, use external tools, and perform reasoning tasks. |
| 22 | [DB-GPT](https://nowfound.app/launch/db-gpt-an-llm-tool-for-database) | Free | 2026-11-21 | open-source agentic AI data assistant for the next generation of AI + Data products. |
| 23 | [Tiny-vLLM](https://nowfound.app/launch/tiny-vllm-high-performance-llm-inference-engine-in-c-and-cud) | Free | 2026-11-20 | Build your own high performance LLM inference engine in C++ and CUDA - a smaller version of vLLM |
| 24 | [Lenz](https://nowfound.app/launch/lenz) | Free plan, paid from $7.99/mo | 2026-09-05 | Independent, multi-model fact-checking API for AI workflows |

## Free tools by topic

- [Free AI tools](https://nowfound.app/free/artificial-intelligence.md) (19,983)
- [Free open-source tools](https://nowfound.app/free/open-source.md) (4,775)
- [Free Development tools](https://nowfound.app/free/development.md) (598)
