AI & Automation · AI Models & Inference
10 paid against 16 free you can run yourself.
Paid against free › AI & Automation › AI Models & Inference › Language Models · AI and models on Repo Radar
Language model APIs give applications direct access to the model makers' own models, paid per token. The paid options range from frontier labs (OpenAI, Anthropic, Google, xAI) to lower-cost providers of strong models (DeepSeek, Moonshot's Kimi, Z.ai's GLM, Alibaba's Qwen, Mistral) and enterprise-focused Cohere. Prices differ by more than a hundred times between the cheapest and most capable models, and most vendors add discounts for cached input and batch jobs. On the open-source side several of the same labs publish open weights (gpt-oss, Qwen, Gemma, Kimi K3) that you can run locally with Ollama or llamafile, fine-tune with LlamaFactory or Transformers, and constrain with Outlines or Instructor. Marks come from vendor pricing pages and project READMEs as of 18 October 2026; nothing was tested hands-on.
Prices and free tiers as checked, dated on each review.
OOpenAI APIOpenAI's platform for building AI products with frontier models and APIs.
free tierCClaude PlatformBuild, deploy, and manage AI agents with Claude models, tools, and infrastructure.
—GGemini APIGoogle's API for Gemini models, from prototyping to production.
free tierMMistral AI PlatformMistral AI platform with Vibe, Studio, and Admin for LLM access and management.
free tierCCohereSecure, scalable AI models and platform for enterprise applications.
$2000free tierGGrok APIGrok API offers models for text, image, video, and voice with usage-based pricing.
free tierDDeepSeek APIDeepSeek API offers OpenAI/Anthropic-compatible language models with reasoning, tool use, and search.
free tierKKimi API PlatformAPI platform for Kimi AI models with pay-as-you-go pricing.
free tierZZ.AI Developer PlatformZ.AI offers GLM model APIs and agent tools for coding and office automation.
free tierAAlibaba Cloud Model StudioOne-stop model service platform for Qwen and third-party models.
free tierFree to use. You run it on your own server.
They do this job.
It downloads and runs AI language models on the computer in front of you.
★ 183kFreeTRhuggingface/transformers🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.
★ 168kFreeLLhiyouga/LlamaFactoryUnified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)
★ 75.6kFreeLLmozilla-ai/llamafileA whole language model and the program that runs it, in one downloadable file.
★ 26.4kFreeGPopenai/gpt-ossgpt-oss-120b and gpt-oss-20b are two open-weight language models by OpenAI
★ 20.5kFreeLILibreTranslate/LibreTranslateSelf-hosted machine-translation API — removes the per-character meter from running a bilingual site.
★ 17.1kFreeOUdottxt-ai/outlinesForces an AI model's answer into a shape you define, while the answer is being produced.
★ 15.9kFreeIN567-labs/instructorMakes a language model return data in a declared shape, and retries automatically when it does not.
★ 14kFreeKIMoonshotAI/Kimi-K3Open Frontier Intelligence
★ 8.9kFreeBEMaartenGr/BERTopicA Python library that groups documents by meaning and labels each group with its distinguishing words.
★ 7.9kFreeGEgoogle-deepmind/gemmaGemma open-weight LLM library, from Google DeepMind
★ 5.8kFreeQWQwenLM/Qwen3.8Qwen3.8 is the large language model series developed by Qwen team, Alibaba Group.
★ 4.3kFreeKEMaartenGr/KeyBERTPulls a few phrases out of a document that say what it is about.
★ 4.2kFreeMOMinishLab/model2vecDistils a sentence model into a static lookup table, so producing embeddings needs no graphics card.
★ 2.2kFreeLIpemistahl/lingua-pyA Python library that detects which of seventy-five languages a piece of text is written in.
★ 1.8kFreeWTsegment-any-text/wtpsplitSplits text into real sentences in 85 languages, even when the writing has no punctuation.
★ 1.4kFreeA second choice: they cover part of this job or are a building block for it.
A private web page for talking to AI models, with accounts and document upload.
★ 156kFreeLLggml-org/llama.cppA C and C++ library and command-line tool for running language models locally.
★ 132kFreeVLvllm-project/vllmA high-throughput and memory-efficient inference and serving engine for LLMs
★ 94.1kFreeNEChatGPTNextWeb/NextChat✨ Zero-config AI chat assistant. No API key needed — sign up and instantly chat with GPT-5, Claude 4, Gemini 2.5, DeepSeek & 100+ top models. Pay-as-you-go saves you more. Available on Web, iOS, macOS, Android, Linux, Windows.
★ 88.9kFreeMEmem0ai/mem0A memory layer that lets an AI application remember a person between separate conversations.
★ 67.2kFreeLIBerriAI/litellmA gateway that puts a hundred or more model providers behind one interface, with spend tracking and limits.
★ 61kFreeCHCherryHQ/cherry-studioAI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs
★ 52.7kFreeLOmudler/LocalAIServes text, vision, speech, image and video models from one self-hosted service that imitates the paid interfaces.
★ 47.8kFreeLILibreChat-AI/LibreChatEnhanced ChatGPT Clone: Features Agents, MCP, Skills, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message search, Code Interpreter, langchain, DALL-E-3, OpenAPI Actions, Functions, Secure Multi-User Auth, Presets, open-source for self-hosting. Active
★ 45.9kFreeJAjanhq/janDownloads and runs language models entirely offline as a desktop application, and opens a local server other programs can use.
★ 45kFreeCHchatboxai/chatboxPowerful AI Client
★ 42.1kFreePRlinshenkx/prompt-optimizerAn AI prompt optimizer for writing better prompts and getting better AI results.
★ 38.6kFreeFRtashfeenahmed/freellmapiOne local address in front of the free tiers of many AI providers.
★ 36.6kFreeLMEleutherAI/lm-evaluation-harnessRuns standard question sets against a language model and reports the scores.
★ 14.3kFreeLLsimonw/llmA command-line tool for sending text to language models, with every exchange logged locally.
★ 12.6kFreeCHhuggingface/chat-uiThe open source codebase powering HuggingChat
★ 11kFreeINxorbitsai/inferenceSwap GPT for any LLM by changing a single line of code. Xinference lets you run open-source, speech, and multimodal models on cloud, on-prem, or your laptop — all through one unified, production-ready inference API.
★ 9.6kFreeDYai-dynamo/dynamoA Datacenter Scale Distributed Inference Serving Framework
★ 8.3kFreeGUguardrails-ai/guardrailsWraps checks around what a language model is asked and what it answers.
★ 7.5kFreeNENVIDIA/NeMo-GuardrailsPuts rules you write around a language model's input and output, and enforces them before the answer reaches anybody.
★ 7.2kFreeMAmagnitudedev/magnitudeIt measures the machine, tunes itself to that hardware, then runs open AI models locally.
★ 6.7kFreeLLAnswerDotAI/llms-txtA proposal for one Markdown file at `/llms.txt` that tells a language model what a site is, plus two tools for working with it.
★ 2.7kFreeLIhuggingface/lightevalRuns standard public tests against a language model and reports the score.
★ 2.6kFreeOTggozad/otermA chat window for AI models that runs in a terminal and keeps past conversations.
★ 2.4kFreeEvery paid tool and free repository we compared, marked against the same features.
✓has itPartlysome plans or extra work—confirmed not available?not confirmed✓the vendor's claim onlypricingClick a mark for its note and source
| Paid · 10 | Free · 10 | |||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Feature | OpenAI API | Claude Platform | Gemini API | Mistral AI Platform | Cohere | Grok API | DeepSeek API | Kimi API Platform | Z.AI Developer Platform | Alibaba Cloud Model Studio | ollama/ollama | mozilla-ai/llamafile | openai/gpt-oss | QwenLM/Qwen3.8 | google-deepmind/gemma | MoonshotAI/Kimi-K3 | huggingface/transformers | hiyouga/LlamaFactory | dottxt-ai/outlines | 567-labs/instructor |
| Starts at | Pay per token; no subscription | Pay per token; no subscription | Pay per token; no subscription | Pay per token; no subscription | From $3.00 per hour per instance (Model Vault) | Pay per token; no subscription | Pay per token; no subscription | Pay per token; no subscription | Pay per token; no subscription | Pay per token; no subscription | MIT | Apache-2.0 | Apache-2.0 | Apache-2.0 | Apache-2.0 | Kimi K3 License | Apache-2.0 | Apache-2.0 | Apache-2.0 | MIT |
| Models | ||||||||||||||||||||
| Frontier model via APIThe vendor's most capable language model is available over its own API. | ||||||||||||||||||||
| Low-cost small modelA small model priced at about $0.20 or less per 1M input tokens. | ||||||||||||||||||||
| Long context (1M tokens)At least one model accepts about a million tokens of context. | ||||||||||||||||||||
| Pricing | ||||||||||||||||||||
| Prompt caching discountCached input tokens cost less than fresh ones. | ||||||||||||||||||||
| Batch discountCheaper rates for jobs that can wait. | ||||||||||||||||||||
| Free tierSome API usage free without paying. | ||||||||||||||||||||
| Customisation | ||||||||||||||||||||
| Fine-tuningTrain your own version of the model. | ||||||||||||||||||||
| Models | ||||||||||||||||||||
| Image, audio or video modelsModels beyond text: images, speech, video or multimodal input. | ||||||||||||||||||||
| Capabilities | ||||||||||||||||||||
| Built-in toolsServer-side tools such as web search, browsing or computer use. | ||||||||||||||||||||
| Compliance | ||||||||||||||||||||
| Data residency optionsChoose the region where requests are processed. | ||||||||||||||||||||
| Ownership | ||||||||||||||||||||
| Open weightsModel weights you can download and run yourself. | ||||||||||||||||||||
| Runs locallyRun language models on your own computer. | ||||||||||||||||||||
| Capabilities | ||||||||||||||||||||
| Structured outputsConstrain output to JSON or a schema. | ||||||||||||||||||||
| Models | ||||||||||||||||||||
| Embedding or rerank modelsModels for search: embeddings or reranking. | ||||||||||||||||||||
| Compliance | ||||||||||||||||||||
| Private or on-prem deploymentDeploy the models in your own cloud or data centre. | ||||||||||||||||||||
| Integration | ||||||||||||||||||||
| OpenAI-compatible APIAccepts code written for the OpenAI API. | ||||||||||||||||||||
| Plan | Price | Billing | Unit | Research allowance | Other limits |
|---|---|---|---|---|---|
| GPT-6 Astra | $10.00 in ($1.00 cached) / $50.00 out | Usage; batch -50% | Per 1M tokens | Not applicable | Data residency +10% |
| GPT-6.1 Sol | $2.00 in ($0.10 cached) / $10.00 out | Usage; batch -50% | Per 1M tokens | Not applicable | Data residency +10% |
| GPT-6 Luna | $0.10 in ($0.01 cached) / $0.50 out | Usage; batch -50% | Per 1M tokens | Not applicable | Data residency +10% |
Checked 2026-10-18 on the pricing page. Prices change; the vendor's page is the final word.
| Plan | Price | Billing | Unit | Research allowance | Other limits |
|---|---|---|---|---|---|
| Claude Opus 5.5 | $8 in / $40 out | Usage; batch -50% | Per 1M tokens | Not applicable | Fast mode available |
| Claude Opus 5 / Opus 4.8 | $10 in / $50 out | Usage; batch -50% | Per 1M tokens | Not applicable | 1M context |
| Claude Haiku 5.5 | Not captured | Usage | Per 1M tokens | Not applicable | Higher price for prompts over 100,000 tokens |
Checked 2026-10-18 on the pricing page. Prices change; the vendor's page is the final word.
| Plan | Price | Billing | Unit | Research allowance | Other limits |
|---|---|---|---|---|---|
| Gemini 4 Argon | $2.00 in (rises to $4.00 on 1 Dec 2026) | Usage; batch -50% | Per 1M tokens | Not applicable | No free tier |
| Gemini 3.8 Flash | $0.75 in / $3.75 out (doubles on 1 Jan 2027) | Usage; batch -50% | Per 1M tokens | Not applicable | Free tier available |
| Free tier | Free of charge | — | Per request limits | Not applicable | Most models |
Checked 2026-10-18 on the pricing page. Prices change; the vendor's page is the final word.
| Plan | Price | Billing | Unit | Research allowance | Other limits |
|---|---|---|---|---|---|
| Mistral Large 4 | $0.68 in (sale, was $1.36) | Usage; batch and priority tiers | Per 1M tokens | Not applicable | Sale price shown |
| Mistral Small 4 | $0.15 in / $0.6 out | Usage | Per 1M tokens | Not applicable | Cached input $0.015 |
| Ministral 3 3B | $0.1 in / $0.1 out | Usage | Per 1M tokens | Not applicable | Cached input $0.01 |
| OCR 4.1 | $4 | Usage | Per 1,000 pages | Not applicable | Document OCR |
Checked 2026-10-18 on the pricing page. Prices change; the vendor's page is the final word.
| Plan | Price | Billing | Unit | Research allowance | Other limits |
|---|---|---|---|---|---|
| Embed 5 Fast (Small) | $3.00/hr | Hourly or $2,000 monthly | Per instance | Not applicable | Model Vault |
| Rerank 4 Pro (Large) | $10.00/hr | Hourly | Per instance | Not applicable | Model Vault |
| North and Compass | Contact sales | Custom | Enterprise | Not applicable | Workplace platform |
Checked 2026-10-18 on the pricing page. Prices change; the vendor's page is the final word.
| Plan | Price | Billing | Unit | Research allowance | Other limits |
|---|---|---|---|---|---|
| Grok 4.7 | $2.00 in / $6.00 out | Usage | Per 1M tokens | Not applicable | Flagship |
| Voice API | $0.08/min speech-to-speech | Usage | Per minute | Not applicable | TTS $15 per 1M chars |
| Imagine API | From $0.02 per image | Usage | Per image | Not applicable | Video from $0.02 per second |
Checked 2026-10-18 on the pricing page. Prices change; the vendor's page is the final word.
| Plan | Price | Billing | Unit | Research allowance | Other limits |
|---|---|---|---|---|---|
| deepseek-flash (V4.1 Flash) | $0.15 in / $0.6 out off-peak; $0.3 / $1.2 peak | Usage | Per 1M tokens | Not applicable | Cache hit $0.003 off-peak |
| deepseek-v4-pro | $0.66 in / $1.98 out off-peak; $1.32 / $3.96 peak | Usage | Per 1M tokens | Not applicable | Cache hit $0.022 off-peak |
| Search and Fetch | $0.75 | Usage | Per 1,000 successful requests | Not applicable | Server-side tools |
Checked 2026-10-18 on the pricing page. Prices change; the vendor's page is the final word.
| Plan | Price | Billing | Unit | Research allowance | Other limits |
|---|---|---|---|---|---|
| kimi-k3 | $3.00 in ($0.30 cached) / $15.00 out | Usage | Per 1M tokens | Not applicable | Cache writes billed by TTL; 1M context |
| kimi-k2.7-code | $0.95 in ($0.19 cached) / $4.00 out | Usage | Per 1M tokens | Not applicable | 262K context |
| kimi-k2.6 | $0.95 in ($0.16 cached) / $4.00 out | Usage | Per 1M tokens | Not applicable | 262K context |
Checked 2026-10-18 on the pricing page. Prices change; the vendor's page is the final word.
| Plan | Price | Billing | Unit | Research allowance | Other limits |
|---|---|---|---|---|---|
| GLM-5.3 | $1.4 in ($0.26 cached) / $4.4 out | Usage | Per 1M tokens | Not applicable | Flagship |
| GLM-5.3-Flash | $0.15 in ($0.03 cached) / $0.50 out | Usage | Per 1M tokens | Not applicable | Low cost |
| Web Search | $0.01 | Usage | Per use | Not applicable | Built-in tool |
Checked 2026-10-18 on the pricing page. Prices change; the vendor's page is the final word.
| Plan | Price | Billing | Unit | Research allowance | Other limits |
|---|---|---|---|---|---|
| Qwen3.7-Max | $2.5 in / $7.5 out | Usage | Per 1M tokens | Not applicable | International region; 1M context |
| Qwen3.8-Max | Not captured | Usage | Per 1M tokens | Not applicable | Context caching discount |
| Batch inference | 50% off | Usage | Per 1M tokens | Not applicable | Supported models |
Checked 2026-10-18 on the pricing page. Prices change; the vendor's page is the final word.
Looked at, and left out of the comparison, with the reason.
What the research could not settle from the vendors' own pages.
One job at a time: the paid tools worth paying for, the free ones you can run yourself, and who should pick which.
No tracking pixels. One click to leave. We use your address to send Stack Radar and nothing else. Privacy.