Offerings

Every active offering: a model as served by one host, with its current canonical price. Sort by clicking a column header; filter with the row below. Prices are USD per 1M tokens unless the unit toggle is set to per 1M characters (an approximation using a measured, blended tokens-per-character ratio).

Showing tier-tagged and vendor-direct models. Show all offerings, including the long tail

More columns: + Vendor + Region + Cache + Batch + Tokenizer - Last change + Source All columns Default columns

ModelHostVariantRegionInputOutputBlended ↑?ContextLast changeConfidence?
GPT-5 miniPerplexity––0.2502.000.688400,000–aggregator
Qwen3-Coder-480B-A35B 2 sourcesDeepInfra––0.4001.600.700262,1442026-08-27 ↕aggregator
GLM-4.6DeepInfraflex+fp4–0.4001.600.700200,000–api
Llama 3.3 70B InstructOpenRouter––0.7100.7100.710128,000–api
Llama 3.3 70B InstructCoreWeave––0.7100.7100.710128,000–api
Llama 3.3 70B Instruct 2 sourcesAzure AI Foundry––0.7100.7100.710128,000–aggregator
Llama 3.3 70B InstructGoogle Vertex AI–us-central10.7200.7200.720128,000–api
Llama 3.3 70B Instruct 3 sourcesGoogle Vertex AI––0.7200.7200.720128,000–api
Llama 3.3 70B Instruct 2 sourcesAWS Bedrock––0.7200.7200.720128,000–aggregator
Llama 3.3 70B InstructAWS Bedrock–us0.7200.7200.720128,000–aggregator
Mistral Large 3 (25.12)OpenRouter––0.5001.500.750262,144–api
Mistral Large 3 (25.12)Mistral AIzdr–0.5001.500.750262,144–api
Llama 3.3 70B Instruct 2 sourcesSambaNova––0.6001.200.750128,000–api
Mistral Large 3 (25.12)Azure AI Foundry––0.5001.500.750262,144–aggregator
Mistral Large 3 (25.12) 2 sourcesAWS Bedrock––0.5001.500.750262,144–aggregator
Mistral Large 3 (25.12) 2 sourcesMistral AI––0.5001.500.750262,1442026-03-02 ↓aggregator
GPT-5 miniAzure AI Foundry–swedencentral0.2752.200.756400,000–api
GPT-5 miniAzure AI Foundry–eu0.2752.200.756400,000–aggregator
GPT-5 miniAzure AI Foundry–us0.2752.200.756400,000–aggregator
GLM-4.6Venicefp4–0.4301.750.760200,000–api
Mistral Large 3 (25.12)AWS Bedrock–ap-southeast-20.5151.550.773262,144–aggregator
Llama 3.3 70B InstructCloudflare Workers AIfp8–0.2932.250.783128,000–api
Qwen3-Coder-480B-A35B 2 sourcesAWS Bedrock––0.4501.800.788262,1442026-09-06 ↑aggregator
Qwen3-Coder-480B-A35BFireworks AI––0.4501.800.788262,144–aggregator
Qwen3-235B-A22B 2 sourcesAlibaba Cloud––0.4551.820.796131,072–api
Mistral Medium 3.1Mistral AI––0.4002.000.800131,072–aggregator
Qwen3-Coder-480B-A35BAWS Bedrock–ap-southeast-20.4641.850.811262,144–aggregator
Qwen3-Coder-480B-A35BAWS Bedrock–ap-southeast-30.4701.870.820262,144–aggregator
Mistral Large 3 (25.12)Mistral AI–eu0.5501.650.825262,144–api
Gemini 2.5 FlashOpenRouter––0.3002.500.8501,048,576–api
MiniMax-M1MiniMax––0.4002.200.8501,000,000–api
Gemini 2.5 Flash 4 sourcesGoogle––0.3002.500.8501,048,576–official
Gemini 2.5 Flash 3 sourcesGoogle Vertex AI––0.3002.500.8501,048,576–api
Gemini 2.5 Flash 2 sourcesDeepInfra––0.3002.500.8501,048,5762025-10-04 ↑official
GLM-4.6OpenRouter––0.5002.000.875200,000–api
GLM-4.6DeepInfrafp4–0.5002.000.875200,000–api
Kimi K2DeepInfra––0.5002.000.875131,072–aggregator
GLM-4.6 3 sourcesDeepInfra––0.5002.000.875200,000–official
DeepSeek V3.2 (chat)Together AI––0.6001.700.875128,000–aggregator
Mistral Large 3 (25.12)AWS Bedrock–ap-south-10.5901.760.883262,144–aggregator
Mistral Large 3 (25.12)AWS Bedrock–ap-northeast-10.6101.820.913262,144–aggregator
Mistral Large 3 (25.12)AWS Bedrock–sa-east-10.6101.820.913262,144–aggregator
Llama 4 MaverickSambaNova––0.6301.800.9231,000,000–aggregator
Qwen3-Coder-480B-A35BAWS Bedrock–ap-south-10.5302.120.928262,144–aggregator
Qwen3-Coder-480B-A35BAWS Bedrock–sa-east-10.5402.160.945262,144–aggregator
Qwen3-Coder-480B-A35BAWS Bedrock–ap-northeast-10.5402.180.950262,144–aggregator
GLM-4.6 3 sourcesNovita AI––0.5502.200.963200,0002026-01-13 ↓api
MiniMax-M1 3 sourcesNovita AI––0.5502.200.9631,000,0002026-01-13 ↓api
DeepSeek V4 ProDeepSeekoff-peak–0.6601.980.9901,048,576–official
GLM-4.6Zhipu AI (Z.ai)fp4–0.6002.201.00200,000–api

Confidence: official = vendor pricing page · api = OpenRouter per-host · aggregator = community list, may lag · unreviewed = automatic extraction awaiting review. When sources disagree the most authoritative wins. Details: /methodology#confidence