Directory
Compare models vs apps — live OpenRouter pricing, no fake benchmarks
Pricing moves this week
OpenAI: gpt-oss-20b (batch)▲ $0.024/M · $0.112Meta: Llama 3.1 8B Instruct▲ $0.05/M · $0.08Mistral: Mistral Small 3▲ $0.05/M · $0.08Nex AGI: Nex-N2.5-Mini▲ $0.025/M · $0.1OpenAI: gpt-oss-120b▲ $0.026/M · $0.096Z.ai: GLM 5.3 Flash▲ $0.029/M · $0.092Z.ai: GLM Flash Latest▲ $0.029/M · $0.092Sao10K: Llama 3 8B Lunaris▲ $0.04/M · $0.05
| Model | Tier | Best for | Input | Context | Status |
|---|---|---|---|---|---|
| Apodex: Apodex 1.1 Mini (free)Apodex | Free | Best for complex research and forecasting; weak at real-time conversational tasks. | Free | 262k | Live |
| DeepSeek: DeepSeek V4 Flash 0731 (free)DeepSeek | Free | Best for coding and reasoning tasks; weak at handling large-scale data processing. | Free | 1049k | Live |
| Dots Studio: Dots3-Note Preview (free)Dots Studio | Free | Best for lightweight applications; weak at handling complex tasks requiring deep contextual understanding. | Free | 512k | Live |
| Gemini (Free tier)Google | Free | Best for quick multimodal Q&A at $0; weak when you need production rate limits or agent loops. | Free | — | AVAILABLE |
| Google: Gemma 4 26B A4B (free)Google | Free | Best for high-quality instruction tasks; weak at handling complex reasoning or extensive contextual understanding. | Free | 262k | Live |
| Google: Gemma 4 31B (free)Google | Free | Best for multimodal tasks combining text and images; weak at handling very long context beyond 256K tokens. | Free | 262k | Live |
| Google: Lyria 3 Clip PreviewGoogle | Free | Best for generating short music clips; weak at producing longer compositions. | Free | 1049k | Live |
| Google: Lyria 3 Pro PreviewGoogle | Free | Best for generating high-quality music tracks; weak at real-time audio processing. | Free | 1049k | Live |
| Meta AI (Llama)Meta | Free | Best for casual chat inside Meta apps; weak as a developer API or offline runtime. | Free | — | AVAILABLE |
| MiniMax: MiniMax M3 (free)MiniMax | Free | Best for multimodal tasks like coding and creative projects; weak at real-time interactive applications. | Free | 1049k | Live |
| MiniMax: MiniMax M3.1 Flash (free)MiniMax | Free | Best for complex agentic tasks and coding; weak at simple conversational queries. | Free | 1000k | Live |
| NVIDIA: Nemotron 3 Super (free)NVIDIA | Free | Best for complex multi-agent applications; weak at simple single-agent tasks. | Free | 262k | Live |
| NVIDIA: Nemotron 3 Ultra (free)NVIDIA | Free | Best for complex reasoning tasks; weak at real-time data processing. | Free | 1000k | Live |
| NVIDIA: Nemotron 3.5 Lightning (free)NVIDIA | Free | Best for high-throughput agentic workloads; weak at general-purpose conversational tasks. | Free | 1000k | Live |
| NVIDIA: Nemotron 3.5 Super VL (free)NVIDIA | Free | Best for multimodal tasks combining text and images; weak at handling purely numerical data. | Free | 1000k | Live |
| Nex AGI: Nex-N2.5-Mini (free)Nex AGI | Free | Best for coding tasks with visual feedback; weak at handling non-technical queries. | Free | 262k | Live |
| Nex AGI: Nex-N2.5-Pro (free)Nex AGI | Free | Best for agentic coding and implementing multi-file projects; weak at handling non-coding tasks or providing detailed explanations. | Free | 262k | Live |
| Poolside: Laguna S 2.1 (free)Poolside | Free | Best for coding assistance; weak at handling complex natural language understanding tasks. | Free | 262k | Live |
| Poolside: Laguna XS 2.1 (free)Poolside | Free | Best for coding assistance; weak at handling complex multi-step reasoning tasks. | Free | 262k | Live |
| Qwen: Qwen3.8 27B (free)Qwen | Free | Best for multimodal interaction and coding; weak at handling highly specialized domain knowledge. | Free | 262k | Live |
| Recursive: Flint (free)Recursive | Free | Best for coding tasks with efficient parameter usage; weak at handling very large datasets or complex reasoning. | Free | 262k | Live |
| Space Bunny AlphaStealth | Free | Best for coding tasks with extensive context; weak at nuanced conversational understanding. | Free | 1000k | Live |
| Thinking Machines: Inkling (free)Thinking Machines | Free | Best for general-purpose reasoning and coding; weak at highly specialized tasks requiring deep domain knowledge. | Free | 1049k | Live |
| Thinking Machines: Inkling Small (free)Thinking Machines | Free | Best for efficient multimodal tasks; weak at handling large-scale data processing. | Free | 1049k | Live |
| Union AlphaStealth | Free | Best for research and coding tasks; weak at casual conversation and unstructured dialogue. | Free | 262k | Live |
| XingChen: Xing4.0 29B A4B (free)XingChen | Free | Best for conversational AI with diverse topics; weak at highly specialized technical queries. | Free | 262k | Live |
| inclusionAI: Ling 3.0 Flash Fin (free)inclusionAI | Free | Best for financial analysis and investment strategies; weak at general conversation and non-financial topics. | Free | 262k | Live |
| inclusionAI: Ling 3.0 Flash Sante (free)inclusionAI | Free | Best for health-related queries; weak at general knowledge outside medical topics. | Free | 262k | Live |
| inclusionAI: Ling 3.0 Flash VL (free)inclusionAI | Free | Best for language tasks with visual context; weak at complex reasoning without visual input. | Free | 262k | Live |
| inclusionAI: Ling 3.1 FlashinclusionAI | Free | Best for complex reasoning tasks; weak at simple data retrieval. | Free | 262k | Live |
Free tokens & offers
Verified free entry points — terms change.
Free modelsOpenRouter
Free models, one API key
Dozens of models listed at $0 for both input and output tokens — rate-limited but genuinely free. Any OpenRouter-compatible client works.
Verified 2026-09 · terms may change
Free tierGoogle AI Studio
Gemini API free tier
Free-of-charge tier for the Gemini API with generous per-minute quotas — a card is required for billing identity, but the free quota isn't billed.
Verified 2026-09 · terms may change
Free tierGroq
Fast inference, free tier
Free tier on GroqCloud for Llama, Mixtral and other open models at very high tokens/second — great for prototyping agents and chat apps.
Verified 2026-09 · terms may change
Free tierMistral
La Plateforme experiment tier
Experiment plan lets you try Mistral models via API for free within a small usage envelope before committing to a paid plan.
Verified 2026-09 · terms may change
Free creditsGitHub
GitHub Models playground
Free API access to GPT, Llama, Mistral and more for GitHub users within daily rate limits, straight from your GitHub account.
Verified 2026-09 · terms may change
Free tierCerebras
Free tier on wafer-scale inference
Free tier for extremely fast inference of open models — useful when latency, not cost, is your bottleneck.
Verified 2026-09 · terms may change