Router DeskIndependent rate desk
OpenRouter pricing, free models, and the API URL
Credits, the $0 router, and per-million rates for DeepSeek, Qwen, Claude, and GLM. Checked 27 Sep 2026.
- Credit purchase
- 5.5%
- Card fee, $0.80 minimum. Crypto is 5%.
- Free router
- $0
- Slug openrouter/free. A negative balance blocks it.
- API base URL
- /api/v1
- https://openrouter.ai/api/v1 for OpenAI clients.
- DeepSeek V4 Pro
- $0.35
- 0423 input per million. 0813 output is $3.50.
- Sonnet 5
- $2 / $10
- Per million input and output tokens.
Account and API
- what-is-openrouterWhat is OpenRouterOpenRouter is one API and one credit balance for hundreds of chat, image, and audio models. Inference is passed through at the provider rate.
- openrouter-pricingOpenRouter pricing and creditsOpenRouter credits are a prepaid dollar balance. Buying credits costs 5.5% ($0.80 minimum). Inference uses each provider’s own token price.
- openrouter-free-modelsOpenRouter free modelsopenrouter/free routes to a $0 model at random. On 27 Sep 2026 the catalog also listed 17 slugs ending in :free, with rate limits.
- openrouter-apiOpenRouter API URLThe OpenRouter base URL is https://openrouter.ai/api/v1. Chat calls go to /chat/completions with a Bearer key.
- openrouter-loginOpenRouter loginOpenRouter login is only on openrouter.ai/sign-in. API keys are created after that, at settings/keys. This site cannot sign you in.
- openrouter-zdrZDR on OpenRouterZDR is Zero Data Retention. On OpenRouter it limits routing to endpoints that do not store prompts. It is separate from the logging opt-in.
Models and prices
- deepseek-api-priceDeepSeek API priceOn 27 Sep 2026 OpenRouter listed DeepSeek V4 Pro 0423 at $0.348 / $0.696 per million tokens, and V4 Pro 0813 at $0.245 / $3.50.
- deepseek-v4-flashDeepSeek V4 Flash priceDeepSeek V4 Flash 0731 was $0.021 / $0.32 per million tokens on 27 Sep 2026. The 0423 slug was $0.047 / $0.094. Neither was the free router.
- openrouter-deepseekDeepSeek on OpenRouterOpenRouter lists several DeepSeek slugs at once: V4 Pro, V4 Flash, older V3 chats, and R1 0528. The slug in your request picks the price.
- deepseek-v3-0324DeepSeek V3 0324DeepSeek V3 0324 still had endpoints on 27 Sep 2026, around $0.24–$0.29 per million input tokens. It is an older chat slug, not V4.
- deepseek-r1-0528DeepSeek R1 0528DeepSeek R1 0528 was still routed on 27 Sep 2026 at about $0.50–$0.70 per million input tokens and $2.15–$2.50 output.
- deepseek-v3-1-terminusDeepSeek V3.1 TerminusDeepSeek V3.1 Terminus was about $0.27 per million input tokens on healthy endpoints on 27 Sep 2026. Two of four providers were not in service.
- qwen3-32bQwen3 32BQwen3 32B is a dense 32.8B model. On 27 Sep 2026 OpenRouter priced qwen/qwen3-32b at $0.08 input and $0.28 output per million tokens.
- qwen3-5-9bQwen 3.5 9BQwen3.5-9B is a 9B multimodal model. OpenRouter listed qwen/qwen3.5-9b at $0.10 input and $0.15 output per million tokens, with a 262K context.
- qwen3-maxQwen3 MaxQwen3.7 Max was $1.475 input and $4.425 output per million tokens on 27 Sep 2026, with a 1M context. The older qwen/qwen3-max slug was $0.78 / $3.90.
- qwen3-235bQwen3 235BQwen3 235B A22B activates 22B of 235B parameters. On 27 Sep 2026 it was $0.455 input and $1.82 output per million tokens.
- qwen3-14bQwen3 14BQwen3 14B is a dense 14.8B model. OpenRouter listed qwen/qwen3-14b at $0.12 input and $0.24 output per million tokens.
- glm-5-2GLM 5.2 APIGLM 5.2 on OpenRouter was $0.650 input and $2.042 output per million tokens on 27 Sep 2026. The slug is z-ai/glm-5.2, with a 1M context.
- ai-model-comparisonAI model comparisonA price and context comparison of DeepSeek V4, Qwen3.7 Max, GLM 5.2, Claude Sonnet 5, GPT-5.1, and Grok 4.5 on OpenRouter, checked 27 Sep 2026.
- best-open-weight-modelsOpen-weight modelsOpen-weight models you can call on OpenRouter in Sep 2026 include DeepSeek V4, Qwen3 32B and 235B, GLM 5.2, and the free Gemma and Nemotron slugs.
- best-embedding-modelsEmbedding modelsOpenRouter ranks embedding models separately from chat. Pick by retrieval score on your corpus, dimension, and price per million tokens.
- gpt-5-1-apiGPT-5.1 APIGPT-5.1 on OpenRouter was $1.25 input and $10 output per million tokens on 27 Sep 2026. The slug is openai/gpt-5.1, with a 400K context.
- gpt-4o-pricingGPT-4o pricingGPT-4o on OpenRouter was $2.50 input and $10 output per million tokens on 27 Sep 2026. The May 2024 snapshot was $5 / $15.
- grok-api-pricingGrok API pricingGrok 4.5 on OpenRouter was $2 input and $6 output per million tokens on 27 Sep 2026, doubling after 200K prompt tokens. The slug is x-ai/grok-4.5.
- grok-4-1-fastGrok 4.1 FastGrok 4.1 Fast is xAI’s tool-calling model with a 2M context on its OpenRouter card. On 27 Sep 2026 it had no live endpoint.
- whisper-large-v3-turboWhisper Large V3 TurboWhisper Large V3 Turbo on OpenRouter is billed per second of audio. On 27 Sep 2026 DeepInfra was $0.000003/sec and Groq was $0.000011/sec.
- seedance-2-0Seedance 2.0 priceSeedance 2.0 is ByteDance video generation. OpenRouter’s page lists from $0.06726 per second at 480p with audio to $0.7776 per second at 4K.
- image-modelsImage routerOpenRouter’s image collection includes Gemini and GPT image slugs. Video such as Seedance is billed per second, not as a chat completion.
- mythomax-l2-13bMythoMax L2 13BMythoMax 13B is a Llama 2 13B fine-tune aimed at description and roleplay. On 27 Sep 2026 it was $0.08 input and $0.11 output, with an 8K context.
- dolphin-3-0Dolphin 3.0Dolphin 3.0 R1 Mistral 24B is an instruct tune for coding, math, and tool calls. On 27 Sep 2026 the slug had a card and no live endpoint.
- opus-4-pricingOpus 4 pricingClaude Opus 4’s card is still on OpenRouter, with no endpoint on 27 Sep 2026. Live Opus rows that day ran from $5 / $25 to $15 / $75 per million tokens.
- claude-haiku-vs-sonnetClaude Haiku vs SonnetHaiku 4.5 is $1 / $5 per million tokens with a 200K context. Sonnet 4.5 is $3 / $15 and up to 1M. Sonnet 5 is $2 / $10.
- claude-sonnet-5-pricingClaude Sonnet 5 pricingClaude Sonnet 5 was $2 input and $10 output per million tokens on 27 Sep 2026. Batch is $1 and $5. The slug is anthropic/claude-sonnet-5.
Apps and agents
- chub-aiChub AIChub AI is a third-party character app listed on OpenRouter. You bring an API key. This site does not host Chub chats or sell a Chub login.
- janitor-aiJanitor AI and OpenRouterJanitor AI can use OpenRouter as an OpenAI-compatible API. The base URL is https://openrouter.ai/api/v1. That is the proxy. It is not a VPN.
- best-ai-for-roleplayAI for roleplayRoleplay on OpenRouter is a model choice: context length, price, and whether the card is written for fiction. Longer chats need a bigger window.
- isekai-zeroIsekai ZeroIsekai Zero is an OpenRouter-listed app for AI adventures with characters. The model bill is still an OpenRouter slug and key.
- darkbloomDarkbloomDarkbloom is a model provider on OpenRouter, not a chat app. On 27 Sep 2026 it listed 8 models, with input rates from about $0.02 to $0.08 per million.
- chutes-aiChutes AIChutes is an inference provider on OpenRouter. On 27 Sep 2026 the provider page listed 6 models, including Qwen and Moonshot rows.
- relaceRelaceRelace on OpenRouter is a small set of coding models. Relace Search was $1 / $3 per million tokens. Relace Apply was $0.85 / $1.25. Both list 256K context.
- lemonadeLemonadeLemonade, listed on OpenRouter, is described as the AI tool for Roblox games. It is an app listing, not a text model.
- best-coding-modelsCoding modelsCoding-model choice on OpenRouter is price, context, and tool-call reliability. DeepSeek V4, Qwen3.7 Max, GLM 5.2, and Sonnet 5 are the slugs to price first.
- best-coding-agentsCoding agentsOpenRouter’s coding-agent apps include Claude Code, OpenCode, and others. The agent is the client. The model is the slug you attach.
- claude-code-openrouterClaude Code on OpenRouterPoint Claude Code at https://openrouter.ai/api, put the OpenRouter key in ANTHROPIC_AUTH_TOKEN, and set ANTHROPIC_API_KEY to an empty string.
- opencode-openrouterOpenCode and OpenRouterOpenCode is an open-source coding agent. Connect OpenRouter with /connect in the terminal, or the same credentials in the desktop app.
- openrouter-mcpOpenRouter MCPThe OpenRouter MCP server is https://mcp.openrouter.ai/mcp. It lets a coding agent read models, prices, and your credits after an OAuth login.
- openclaw-modelsModels for OpenClawOpenClaw is a coding agent with an OpenRouter model collection. Pick a slug from that collection, then judge it on context, tools, and output price.
Questions
What is OpenRouter?
A gateway with one API key and a dollar credit balance for many models. Inference is billed at the provider’s token price.
How much do OpenRouter credits cost?
Credits are dollars. Buying them on a card adds a 5.5 percent fee, with an $0.80 minimum.
What is the API base URL?
https://openrouter.ai/api/v1. Claude Code uses https://openrouter.ai/api without the /v1 suffix.
Are there free models?
Yes. openrouter/free picks one at random, and other slugs end in :free. They stop working if the account balance is negative.