以下计算均按照下表计算得出:Input TokensOutput TokensAPI Calls10001000100ProviderModelContextInput/1k TokensOutput/1k TokensPer CallTotalChat/Completion ModelsOpenAI / AzureGPT-3.5 Turbo16K$0.0005$0.0015$0.0080$0.80OpenAI / AzureGPT-4 Turbo128K$0.01$0.03$0.1600$16.00OpenAI / AzureGPT-48K$0.03$0.06$0.3300$33.00AnthropicClaude Instant100K$0.0008$0.0024$0.0128$1.28AnthropicClaude 2.1200K$0.008$0.024$0.1280$12.80AnthropicClaude 3 Haiku200K$0.00025$0.00125$0.0065$0.65AnthropicClaude 3 Sonnet200K$0.003$0.015$0.0780$7.80AnthropicClaude 3 Opus200K$0.015$0.075$0.3900$39.00Meta (via Anyscale)Llama 2 70b4K$0.001$0.001$0.0060$0.60GoogleGemini 1.0 Pro32K$0.0005$0.0015$0.0080$0.80CohereCommand4K$0.01$0.02$0.1100$11.00CohereCommand R128K IN/4K OUT$0.0005$0.0015$0.0080$0.80CohereCommand R+128K$0.003$0.015$0.0780$7.80Mistral AI (via Anyscale)Mixtral 8x7B32K$0.0005$0.0005$0.0030$0.30Mistral AIMistral Small32K$0.002$0.006$0.0320$3.20Mistral AIMistral Large32K$0.008$0.024$0.1280$12.80DataBricksDBRX32K$0.00225$0.00675$0.0360$3.60Fine-tuning modelsOpenAIGPT-3.5 Turbo4K$0.012$0.016$0.0920$9.20GooglePaLM 28K$0.002$0.002$0.0120$1.20Embedding modelsOpenAI / AzureAda v2$0.0001$0.0001$0.01GooglePaLM 2$0.0004$0.0004$0.04CohereEmbed$0.0001$0.0001$0.01