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

ModelVendorHostVariantInputOutputBlended ↑?Cache readCache writeBatch inBatch outContextTokenizerLast changeConfidence?Source
Qwen3 Coder 480B A35B Instructalibabagoogle-vertex–0.220.06631.800.540.620.190.02200.0066–n/a0.110.03320.900.27262144qwen2–aggregatorlitellm
Llama 3.3 70B Instructmetagroqunknown0.590.180.790.240.640.190.300.0888–n/a–n/a–n/a131072llama3–apiopenrouter
Llama 3.3 70B Instructmetagroq–0.590.180.790.240.640.190.300.0888–n/a–n/a–n/a131072llama3–apiopenrouter
Qwen3 Coder 480B A35B Instructalibabanovita–0.380.111.550.470.670.20–n/a–n/a–n/a–n/a262144qwen2–aggregatorlitellm
GPT-5 Miniopenaiopenrouter–0.250.07042.000.560.690.190.02500.0070–n/a0.130.03521.000.28400000o200k–apiopenrouter
GPT-5 Miniopenaiazureunknown0.250.07042.000.560.690.190.03000.0085–n/a–n/a–n/a400000o200k–apiopenrouter
GPT-5 Miniopenaiopenaiunknown0.250.07042.000.560.690.190.02500.0070–n/a–n/a–n/a400000o200k–apiopenrouter
GPT-5 Miniopenaiazure–0.250.07042.000.560.690.190.02500.0070–n/a0.130.03521.000.28400000o200k–aggregatorlitellm
GPT-5 Miniopenaiazure–0.250.07042.000.560.690.190.02500.0070–n/a0.130.03521.000.28400000o200k–aggregatorlitellm
GPT-5 Miniopenaiopenai–0.250.07042.000.560.690.190.02500.0070–n/a0.130.03521.000.28400000o200k–aggregatorlitellm
GPT-5 Miniopenaiopenai–0.250.07042.000.560.690.190.02500.0070–n/a0.130.03521.000.28400000o200k–aggregatorlitellm
GPT-5 Miniopenaiperplexity–0.250.07042.000.560.690.190.02500.0070–n/a–n/a–n/a400000o200k–aggregatorlitellm
GPT-5 Miniopenaiazure–0.250.07042.000.560.690.190.03000.0085–n/a–n/a–n/a400000o200k–apiopenrouter
GPT-5 Miniopenaiopenai–0.250.07042.000.560.690.190.02500.0070–n/a–n/a–n/a400000o200k–apiopenrouter
Qwen3 Coder 480B A35B Instructalibabadeepinfra–0.400.121.600.480.700.21–n/a–n/a–n/a–n/a262144qwen2–aggregatorlitellm
Llama 3.3 70B Instructmetacoreweave–0.710.210.710.210.710.210.710.21–n/a–n/a–n/a131072llama3–apiopenrouter
Llama 3.3 70B Instructmetaazure–0.710.210.710.210.710.21–n/a–n/a–n/a–n/a131072llama3–aggregatorlitellm
Llama 3.3 70B Instructmetaazure–0.710.210.710.210.710.21–n/a–n/a–n/a–n/a131072llama3–aggregatormodelsdev
Llama 3.3 70B Instructmetagoogle-vertexunknown0.720.220.720.220.720.22–n/a–n/a–n/a–n/a131072llama3–apiopenrouter
Llama 3.3 70B Instructmetagoogle-vertexunknown0.720.220.720.220.720.22–n/a–n/a–n/a–n/a131072llama3–apiopenrouter
Llama 3.3 70B Instructmetaaws-bedrock–0.720.220.720.220.720.22–n/a–n/a–n/a–n/a131072llama3–aggregatorlitellm
Llama 3.3 70B Instructmetaaws-bedrock–0.720.220.720.220.720.22–n/a–n/a–n/a–n/a131072llama3–aggregatorlitellm
Llama 3.3 70B Instructmetagoogle-vertex–0.720.220.720.220.720.22–n/a–n/a0.360.110.360.11131072llama3–aggregatorlitellm
Llama 3.3 70B Instructmetagoogle-vertex–0.720.220.720.220.720.22–n/a–n/a–n/a–n/a131072llama3–apiopenrouter
Llama 3.3 70B Instructmetagoogle-vertex–0.720.220.720.220.720.22–n/a–n/a–n/a–n/a131072llama3–apiopenrouter
Qwen3 235B A22B Instruct 2507alibabaopenrouter–0.230.06942.300.690.750.23–n/a–n/a–n/a–n/a131072qwen2–apiopenrouter
Qwen3 235B A22B Instruct 2507alibabaalibabaunknown0.230.06942.300.690.750.23–n/a–n/a–n/a–n/a131072qwen2–apiopenrouter
Qwen3 235B A22B Instruct 2507alibabaalibaba–0.230.06942.300.690.750.23–n/a–n/a–n/a–n/a131072qwen2–apiopenrouter
Mistral Large (latest)mistralmistral–0.500.181.500.530.750.270.05000.0178–n/a–n/a–n/a262144mistral–aggregatorlitellm
Llama 3.3 70B Instructmetasambanova–0.600.181.200.360.750.23–n/a–n/a–n/a–n/a131072llama3–aggregatorlitellm
GPT-5 Miniopenaiazureunknown0.280.07752.200.620.760.210.03300.0093–n/a–n/a–n/a400000o200k–apiopenrouter
GPT-5 Miniopenaiazure–0.280.07752.200.620.760.210.03300.0093–n/a–n/a–n/a400000o200k–apiopenrouter
GLM 4.6zhipuopenrouter–0.43n/a1.75n/a0.76n/a0.0800n/a–n/a–n/a–n/a204800––apiopenrouter
GLM 4.6zhipuvenicefp40.43n/a1.75n/a0.76n/a0.0800n/a–n/a–n/a–n/a204800––apiopenrouter
Llama 3.3 70B Instructmetacloudflarefp80.290.08822.250.680.780.24–n/a–n/a–n/a–n/a131072llama3–apiopenrouter
Qwen3 Coder 480B A35B Instructalibabaaws-bedrock–0.450.141.800.540.790.24–n/a–n/a–n/a–n/a262144qwen2–aggregatorlitellm
Qwen3 Coder 480B A35B Instructalibabafireworks–0.450.141.800.540.790.24–n/a–n/a–n/a–n/a262144qwen2–aggregatorlitellm
Qwen3 Coder 480B A35B Instructalibabaaws-bedrock–0.450.141.800.540.790.24–n/a–n/a–n/a–n/a262144qwen2–aggregatorlitellm
Qwen3 235B A22B Instruct 2507alibabaopenrouter–0.460.141.820.550.800.24–n/a–n/a–n/a–n/a131072qwen2–apiopenrouter
Qwen3 235B A22B Instruct 2507alibabaalibabaunknown0.460.141.820.550.800.24–n/a–n/a–n/a–n/a131072qwen2–apiopenrouter
Qwen3 235B A22B Instruct 2507alibabaalibaba–0.460.141.820.550.800.24–n/a–n/a–n/a–n/a131072qwen2–apiopenrouter
Mistral Medium (latest)mistralmistral–0.400.142.000.710.800.28–n/a–n/a–n/a–n/a262144mistral–aggregatormodelsdev
Gemini 2.5 Flashgoogleopenrouter–0.30n/a2.50n/a0.85n/a0.0300n/a0.0833n/a0.15n/a1.25n/a1048576gemini–apiopenrouter
MiniMax M1minimaxminimaxunknown0.40n/a2.20n/a0.85n/a–n/a–n/a–n/a–n/a1000000––apiopenrouter
Gemini 2.5 Flashgooglegoogle-vertexunknown0.30n/a2.50n/a0.85n/a0.0300n/a0.0833n/a–n/a–n/a1048576gemini–apiopenrouter
Gemini 2.5 Flashgooglegoogle-vertexunknown0.30n/a2.50n/a0.85n/a0.0300n/a0.0833n/a–n/a–n/a1048576gemini–apiopenrouter
Gemini 2.5 Flashgooglegoogleunknown0.30n/a2.50n/a0.85n/a0.0300n/a0.0833n/a–n/a–n/a1048576gemini–apiopenrouter
Gemini 2.5 Flashgooglegoogle-vertexunknown0.30n/a2.50n/a0.85n/a0.0300n/a0.0833n/a–n/a–n/a1048576gemini–apiopenrouter
Gemini 2.5 Flashgoogledeepinfra–0.30n/a2.50n/a0.85n/a–n/a–n/a–n/a–n/a1048576gemini–aggregatorlitellm
Gemini 2.5 Flashgooglegoogle-vertex–0.30n/a2.50n/a0.85n/a0.0300n/a–n/a0.15n/a1.25n/a1048576gemini–aggregatorlitellm

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