> ## Documentation Index > Fetch the complete documentation index at: https://prices.voicegateway.dev/llms.txt > Use this file to discover all available pages before exploring further. # LLM pricing > 1095 LLM model(s) from 31 provider(s), largest catalogs first. 1095 LLM model(s) from 31 provider(s). Priced from `input_mtok` and `output_mtok`, US dollars per 1,000,000 tokens. Pick a vendor in the sidebar to see that vendor alone. LLM rows carry no per-model verification status. Unlike voice rates, they are cross-checked against external aggregators rather than against one vendor page each. [How Fresh?](/how-fresh) explains the difference. | Provider | Models | Cheapest input \$ / Mtok | Cheapest output \$ / Mtok | | ----------------------------------------------------------- | -----: | -----------------------: | ------------------------: | | [OpenRouter](/llm/openrouter) | 461 | \$0.005 | \$0.01 | | [Together AI](/llm/together) | 72 | \$0.1 | \$0.1 | | [AWS Bedrock](/llm/aws) | 69 | \$0.035 | \$0.04 | | [OpenAI](/llm/openai) | 63 | \$0.02 | \$0.4 | | [HuggingFace (novita)](/llm/huggingface-novita) | 61 | \$0.02 | \$0.04 | | [Novita](/llm/novita) | 34 | \$0.02 | \$0.02 | | [LiveKit Inference](/llm/livekit) | 33 | \$0.05 | \$0.4 | | [Google](/llm/google) | 31 | \$0.0375 | \$0.15 | | [Groq](/llm/groq) | 29 | \$0.04 | \$0.04 | | [HuggingFace (nebius)](/llm/huggingface-nebius) | 26 | \$0.02 | \$0.06 | | [HuggingFace (together)](/llm/huggingface-together) | 23 | \$0.02 | \$0.04 | | [HuggingFace (nscale)](/llm/huggingface-nscale) | 20 | \$0.01 | \$0.03 | | [Anthropic](/llm/anthropic) | 18 | \$0.25 | \$1.25 | | [Microsoft Azure](/llm/azure) | 18 | \$0.02 | \$0.1 | | [Mistral](/llm/mistral) | 18 | \$0.04 | \$0.04 | | [OVHcloud AI Endpoints](/llm/ovhcloud) | 15 | \$0.01 | \$0.11 | | [Fireworks](/llm/fireworks) | 13 | \$0.07 | \$0.1 | | [HuggingFace (hyperbolic)](/llm/huggingface-hyperbolic) | 12 | \$0.1 | \$0.1 | | [X AI](/llm/x-ai) | 12 | \$0.2 | \$0.5 | | [MoonshotAi](/llm/moonshotai) | 9 | \$0.2 | \$2 | | [HuggingFace (publicai)](/llm/huggingface-publicai) | 8 | \$0.1 | \$0.2 | | [HuggingFace (sambanova)](/llm/huggingface-sambanova) | 8 | \$0.1 | \$0.2 | | [Perplexity](/llm/perplexity) | 8 | \$0.2 | \$0.2 | | [HuggingFace (ovhcloud)](/llm/huggingface-ovhcloud) | 7 | \$0.05 | \$0.11 | | [Cohere](/llm/cohere) | 6 | \$0.0375 | \$0.15 | | [HuggingFace (groq)](/llm/huggingface-groq) | 5 | \$0.1 | \$0.34 | | [Avian](/llm/avian) | 4 | \$0.1 | \$0.1 | | [Cerebras](/llm/cerebras) | 4 | \$0.1 | \$0.1 | | [Deepseek](/llm/deepseek) | 4 | \$0.135 | \$0.28 | | [HuggingFace (fireworks-ai)](/llm/huggingface-fireworks-ai) | 3 | \$0.05 | \$0.2 | | [HuggingFace (cerebras)](/llm/huggingface-cerebras) | 1 | \$0.1 | \$0.1 | ## Via LiveKit Inference What the same model costs going direct to the vendor versus through the gateway. Scale is the discounted tier. | Model | Name | Direct (input \$ / Mtok) | LiveKit (input \$ / Mtok) | Scale (input \$ / Mtok) | vs direct | | --------------------------------- | ----------------------- | -----------------------: | ------------------------: | ----------------------: | --------: | | `google/gemini-2.5-pro` | Gemini 2.5 Pro | \$1.25 | \$2.5 | - | +100.0% | | `google/gemini-3.1-pro-preview` | Gemini 3.1 Pro | \$2 | \$4 | - | +100.0% | | `openai/gpt-5.4` | GPT-5.4 | \$2.5 | \$5 | - | +100.0% | | `openai/gpt-5.5` | GPT-5.5 | \$5 | \$10 | - | +100.0% | | `google/gemini-2.5-flash` | Gemini 2.5 Flash | \$0.3 | \$0.3 | - | same | | `google/gemini-2.5-flash-lite` | Gemini 2.5 Flash-Lite | \$0.1 | \$0.1 | - | same | | `google/gemini-3-flash-preview` | Gemini 3 Flash | \$0.5 | \$0.5 | - | same | | `google/gemini-3.1-flash-lite` | Gemini 3.1 Flash Lite | \$0.25 | \$0.25 | - | same | | `google/gemini-3.5-flash` | Gemini 3.5 Flash | \$1.5 | \$1.5 | - | same | | `moonshotai/kimi-k2.5` | Kimi K2.5 | \$0.6 | \$0.6 | - | same | | `openai/gpt-4.1` | GPT-4.1 | \$2 | \$2 | - | same | | `openai/gpt-4.1-mini` | GPT-4.1 mini | \$0.4 | \$0.4 | - | same | | `openai/gpt-4.1-nano` | GPT-4.1 nano | \$0.1 | \$0.1 | - | same | | `openai/gpt-4o` | GPT-4o | \$2.5 | \$2.5 | - | same | | `openai/gpt-4o-mini` | GPT-4o mini | \$0.15 | \$0.15 | - | same | | `openai/gpt-5` | GPT-5 | \$1.25 | \$1.25 | - | same | | `openai/gpt-5-mini` | GPT-5 mini | \$0.25 | \$0.25 | - | same | | `openai/gpt-5-nano` | GPT-5 nano | \$0.05 | \$0.05 | - | same | | `openai/gpt-5.1` | GPT-5.1 | \$1.25 | \$1.25 | - | same | | `openai/gpt-5.1-chat-latest` | GPT-5.1 Chat | \$1.25 | \$1.25 | - | same | | `openai/gpt-5.2` | GPT-5.2 | \$1.75 | \$1.75 | - | same | | `openai/gpt-5.2-chat-latest` | GPT-5.2 Chat | \$1.75 | \$1.75 | - | same | | `openai/gpt-5.3-chat-latest` | GPT-5.3 Chat | \$1.75 | \$1.75 | - | same | | `openai/gpt-5.4-mini` | GPT-5.4 mini | \$0.75 | \$0.75 | - | same | | `openai/gpt-5.4-nano` | GPT-5.4 nano | \$0.2 | \$0.2 | - | same | | `xai/grok-4-1-fast-non-reasoning` | Grok 4.1 Fast | \$0.2 | \$0.2 | - | same | | `xai/grok-4-1-fast-reasoning` | Grok 4.1 Fast Reasoning | \$0.2 | \$0.2 | - | same | 6 further LiveKit LLM model(s) have no direct-vendor entry to compare against.