TyTo Logo
TyTo AI

Finansiell Analys av AI-Driftsättning

Ett interaktivt verktyg för att jämföra kostnader och prestanda mellan att köpa hårdvara, hyra serverkraft eller använda ett API.

Kostnadstabell för API-modeller

Jämför kostnader mellan olika API-modeller baserat på ditt angivna scenario. Klicka på kolumnrubriker för att sortera.

Modell
Företag
Input-kostnad
Output-kostnad
Kostnad / anrop
Total Dagskostnad
Total (30 dagar)
Llama 3.1 8B Instant 128kGroq$0.000050$0.000080$0.000130$0.0130$0.3900
Llama 3 8B 8kGroq$0.000050$0.000080$0.000130$0.0130$0.3900
Mistral Small 3.2 24Bopenrouter$0.000050$0.000100$0.000150$0.0150$0.4500
Gemma 3 12Bopenrouter$0.000050$0.000100$0.000150$0.0150$0.4500
Qwen3 8Bopenrouter$0.000035$0.000138$0.000173$0.0173$0.5190
Phi 4 14Bopenrouter$0.000070$0.000140$0.000210$0.0210$0.6300
Llama 3.3 70Bopenrouter$0.000050$0.000190$0.000240$0.0240$0.7200
Gemma 3 27Bopenrouter$0.000100$0.000180$0.000280$0.0280$0.8400
Qwen3 14Bopenrouter$0.000060$0.000240$0.000300$0.0300$0.9000
DeepSeek: R1 Distill Qwen 32Bopenrouter$0.000120$0.000180$0.000300$0.0300$0.9000
Mistral 7BAWS Bedrock$0.000150$0.000200$0.000350$0.0350$1.05
gemini 1.5 flash-8B > 128kGoogle$0.000100$0.000300$0.000400$0.0400$1.20
gemini 2.0 flash-liteGoogle$0.000100$0.000300$0.000400$0.0400$1.20
Qwen3 32Bopenrouter$0.000100$0.000300$0.000400$0.0400$1.20
Gemma 2 9B 8kGroq$0.000200$0.000200$0.000400$0.0400$1.20
Llama 4 Scout (17Bx16E)Groq$0.000110$0.000340$0.000450$0.0450$1.35
gpt-4.1-nanoOpenAI$0.000100$0.000400$0.000500$0.0500$1.50
gemini 2.0 flash textGoogle$0.000100$0.000400$0.000500$0.0500$1.50
GPT-4.1-nano (2025-04-14)Azure$0.000100$0.000400$0.000500$0.0500$1.50
DeepSeek-V3-0324 （UTC 16:30-00:30）DeepSeek$0.000035$0.000550$0.000585$0.0585$1.76
DeepSeek-R1-0528 （UTC 16:30-00:30）DeepSeek$0.000035$0.000550$0.000585$0.0585$1.76
Qwen QwQ 32B (Preview) 128kGroq$0.000290$0.000390$0.000680$0.0680$2.04
gemini 1.5 flash > 128kGoogle$0.000100$0.000600$0.000700$0.0700$2.10
Qwen3 235Bopenrouter$0.000130$0.000600$0.000730$0.0730$2.19
gpt-4o-miniOpenAI$0.000200$0.000600$0.000800$0.0800$2.40
grok-3-miniGrok (xAI)$0.000300$0.000500$0.000800$0.0800$2.40
Llama 4 Maverick (17Bx128E)Groq$0.000200$0.000600$0.000800$0.0800$2.40
Mistral 7B Instruct v0.3openrouter$0.000280$0.000540$0.000820$0.0820$2.46
Scout 17BAWS Bedrock$0.000170$0.000660$0.000830$0.0830$2.49
Qwen3 32B 131kGroq$0.000290$0.000590$0.000880$0.0880$2.64
Mixtral 8×7BAWS Bedrock$0.000450$0.000700$0.001150$0.1150$3.45
DeepSeek-V3-0324 （UTC 00:30-16:30）DeepSeek$0.000070$0.001100$0.001170$0.1170$3.51
Maverick 17BAWS Bedrock$0.000240$0.000970$0.001210$0.1210$3.63
Llama 3.3 70B Versatile 128kGroq$0.000590$0.000790$0.001380$0.1380$4.14
Llama 3 70B 8kGroq$0.000590$0.000790$0.001380$0.1380$4.14
Claude Haiku 3Anthropic$0.000300$0.001300$0.001600$0.1600$4.80
Llama 3.1 405Bopenrouter$0.000800$0.000800$0.001600$0.1600$4.80
DeepSeek R1 Distill Llama 70BGroq$0.000750$0.000990$0.001740$0.1740$5.22
gemini 2.0 flash live textGoogle$0.000300$0.001500$0.001800$0.1800$5.40
Dolphin 2.9.2 Mixtral 8x22Bopenrouter$0.000900$0.000900$0.001800$0.1800$5.40
gpt-4.1-miniOpenAI$0.000400$0.001600$0.002000$0.2000$6.00
GPT-3.5-Turbo-0125 (16k)Azure$0.000500$0.001500$0.002000$0.2000$6.00
GPT-4.1-mini (2025-04-14)Azure$0.000400$0.001600$0.002000$0.2000$6.00
DeepSeek-R1-0528 （UTC 00:30-16:30）DeepSeek$0.000140$0.002190$0.002330$0.2330$6.99
NVIDIA: Llama 3.1 Nemotron Ultra 253B v1openrouter$0.000600$0.001800$0.002400$0.2400$7.20
gemini 2.5 flash live textGoogle$0.000500$0.002000$0.002500$0.2500$7.50
DeepSeek: R1 0528 671Bopenrouter$0.000500$0.002150$0.002650$0.2650$7.95
gemini 2.5 flash textGoogle$0.000300$0.002500$0.002800$0.2800$8.40
GPT-3.5-Turbo-1106 (16k)Azure$0.001000$0.002000$0.003000$0.3000$9.00
Mistral Small (24.02)AWS Bedrock$0.001000$0.003000$0.004000$0.4000$12.00
grok-3-mini-fastGrok (xAI)$0.000600$0.004000$0.004600$0.4600$13.80
Claude Haiku 3.5Anthropic$0.000800$0.004000$0.004800$0.4800$14.40
o4-miniOpenAI$0.001100$0.004400$0.005500$0.5500$16.50
o3-miniOpenAI$0.001100$0.004400$0.005500$0.5500$16.50
o1-miniOpenAI$0.001100$0.004400$0.005500$0.5500$16.50
o4-mini (2025-04-16)Azure$0.001100$0.004400$0.005500$0.5500$16.50
gemini 1.5 pro ≤ 128kGoogle$0.001300$0.005000$0.006300$0.6300$18.90
DeepSeek-R1AWS Bedrock$0.001350$0.005400$0.006750$0.6750$20.25
gpt-4.1OpenAI$0.002000$0.008000$0.0100$1.00$30.00
o3OpenAI$0.002000$0.008000$0.0100$1.00$30.00
GPT-4.1 (2025-04-14)Azure$0.002000$0.008000$0.0100$1.00$30.00
gemini 2.5 pro ≤ 200kGoogle$0.001300$0.0100$0.0113$1.13$33.90
grok-2-1212Grok (xAI)$0.002000$0.0100$0.0120$1.20$36.00
gpt-4oOpenAI$0.002500$0.0100$0.0125$1.25$37.50
gemini 1.5 pro > 128kGoogle$0.002500$0.0100$0.0125$1.25$37.50
Mistral Large (24.02)AWS Bedrock$0.004000$0.0120$0.0160$1.60$48.00
gemini 2.5 pro > 200kGoogle$0.002500$0.0150$0.0175$1.75$52.50
Claude Sonnet 4Anthropic$0.003000$0.0150$0.0180$1.80$54.00
Claude Sonnet 3.7Anthropic$0.003000$0.0150$0.0180$1.80$54.00
Claude Sonnet 3.5Anthropic$0.003000$0.0150$0.0180$1.80$54.00
grok-3Grok (xAI)$0.003000$0.0150$0.0180$1.80$54.00
Claude Sonnet 4AWS Bedrock$0.003000$0.0150$0.0180$1.80$54.00
Claude 3.7 SonnetAWS Bedrock$0.003000$0.0150$0.0180$1.80$54.00
Claude 3.5 Sonnet v2AWS Bedrock$0.003000$0.0150$0.0180$1.80$54.00
grok-3-fastGrok (xAI)$0.005000$0.0250$0.0300$3.00$90.00
GPT-4-Turbo (128k)Azure$0.0100$0.0300$0.0400$4.00$120.00
o3 (2025-04-16)Azure$0.0100$0.0400$0.0500$5.00$150.00
o1OpenAI$0.0150$0.0600$0.0750$7.50$225.00
Claude Opus 4Anthropic$0.0150$0.0750$0.0900$9.00$270.00
Claude Opus 3Anthropic$0.0150$0.0750$0.0900$9.00$270.00
GPT-4 (8k)Azure$0.0300$0.0600$0.0900$9.00$270.00
o3-proOpenAI$0.0200$0.0800$0.1000$10.00$300.00
GPT-4 (32k)Azure$0.0600$0.1200$0.1800$18.00$540.00
gpt-4.5-previewOpenAI$0.0750$0.1500$0.2250$22.50$675.00
o1-proOpenAI$0.1500$0.6000$0.7500$75.00$2250.00

Analys av Hyra GPU-server

Jämför och analysera kostnader och prestanda för att hyra dedikerad GPU-hårdvara.

GPU-serverpriser

Timpriser för att hyra dedikerade GPU-servrar från olika leverantörer. Använd kontrollerna för att söka och sortera.

Företag
GPU-modell
VRAM (GB)
Pris ($/h)
Vast.aiRTX A200012$0.0400
Vast.aiGTX 10808$0.0400
Vast.aiGTX 1070 Ti8$0.0400
Vast.aiRTX 30708$0.0500
Vast.aiRTX 20606$0.0500
Vast.aiGTX 1660 S6$0.0500
Vast.aiRTX 306012$0.0600
Vast.aiGTX 10708$0.0600
Vast.aiGTX 1080 Ti11$0.0600
Vast.aiRTX 3060 Ti8$0.0700
Vast.aiRTX 40608$0.0700
Vast.aiRTX 2080 Ti11$0.0800
Vast.aiRTX 3070 Ti8$0.0800
Vast.aiRTX 507012$0.0900
Vast.aiRTX A400016$0.1000
Vast.aiRTX 308010$0.1000
Vast.aiRTX 2060S8$0.1000
Vast.aiRTX 4060 Ti16$0.1200
Vast.aiRTX 5060 Ti16$0.1200
Vast.aiRTX 407012$0.1200
Vast.aiRTX 3080 Ti12$0.1200
Vast.aiRTX 5070 Ti16$0.1300
Vast.aiRTX 4070 Ti12$0.1300
Vast.aiRTX 4070S Ti16$0.1600
Vast.aiRTX 508016$0.1600
Vast.aiRTX 4070S12$0.1600
Vast.aiRTX 309024$0.1700
Vast.aiRTX A500024$0.2000
Vast.aiRTX 4080S16$0.2000
Vast.aiRTX 3090 Ti24$0.2000
Vast.aiTesla V10032$0.2300
RunPodRTX A500024$0.2600
Vast.aiQ RTX 800048$0.3200
Vast.aiRTX 409024$0.3500
RunPodA4048$0.4000
RunPodL424$0.4300
Vast.aiRTX 509032$0.4300
Vast.aiRTX A600048$0.4500
RunPodRTX 309024$0.4600
Vast.aiA4048$0.4600
RunPodRTX A600048$0.4900
Vast.aiRTX 6000ADA48$0.6500
Vast.aiL40S48$0.6800
RunPodRTX 409024$0.6900
RunPodRTX 6000 Ada48$0.7700
Vast.aiA100 SXM480$0.8200
RunPodL40S48$0.8600
Vast.aiA100 PCIE40$0.8700
RunPodRTX 509032$0.8900
RunPodL4048$0.9900
RunPodA100 PCIe80$1.6400
RunPodA100 SXM80$1.7400
Vast.aiH100 SXM80$1.8700
RunPodH100 PCIe80$2.1900
Vast.aiH100 NVL188$2.2700
RunPodH100 NVL94$2.7900
RunPodH100 SXM80$2.9900
Vast.aiH200141$3.0600
RunPodH200 SXM141$3.9900
RunPodB200180$6.3900

Analys av Köpa GPU

Jämför och analysera kostnader och prestanda för att köpa dedikerad GPU-hårdvara.

Priser för att Köpa GPU

Rekommenderade priser (MSRP) för olika grafikkort och enterprise-servrar.

Underkategori
GPU-modell
VRAM (GB)
MSRP ($)
Lokala GPUGeForce GTX 1660 Super6$229
Lokala GPUGeForce RTX 3050 8GB8$249
Lokala GPUGeForce RTX 50508$249
Lokala GPUGeForce RTX 4060 8GB8$299
Lokala GPUGeForce RTX 50608$299
Lokala GPUGeForce RTX 3060 8GB8$300
Lokala GPUGeForce RTX 3060 12GB12$329
Lokala GPUGeForce RTX 5060 Ti 8GB8$379
Lokala GPUGeForce RTX 2060 Super8$399
Lokala GPUGeForce RTX 3060 Ti 8GB8$399
Lokala GPUGeForce RTX 4060 Ti 8GB8$399
Lokala GPUGeForce RTX 5060 Ti 16GB16$429
Lokala GPUGeForce RTX 4060 Ti 16GB16$499
Lokala GPUGeForce RTX 4070 12GB12$549
Lokala GPUGeForce RTX 507012$549
Lokala GPUGeForce RTX 4070 Super12$599
Lokala GPUGeForce RTX 5070 Super18$650
Lokala GPUGeForce RTX 5070 Ti16$749
Lokala GPUGeForce RTX 4070 Ti12$799
Lokala GPUGeForce RTX 4070 Ti Super16$799
Lokala GPUGeForce RTX 4080 Super16$999
Lokala GPUGeForce RTX 508016$999
Lokala GPUGeForce RTX 3080 Ti12$1,199
Lokala GPUGeForce RTX 408016$1,199
Lokala GPURTX A400016$1,200
Lokala GPUGeForce RTX 309024$1,499
Lokala GPUGeForce RTX 5080 Super*24$1,500
Lokala GPUGeForce RTX 409024$1,599
Lokala GPUGeForce RTX 3090 Ti24$1,999
Lokala GPUGeForce RTX 509032$1,999
Lokala GPURTX A500024$2,300
Lokala GPUAMD Radeon Pro W790032$2,500
Lokala GPUQuadro RTX 800048$3,500
Lokala GPUAMD Radeon Pro W790048$3,500
Lokala GPURTX A600048$5,100
Enterprise GPUL40S48$7,500
Lokala GPURTX Pro 6000 Blackwell48$8,500
Enterprise GPUA100 40GB PCIe40$10,000
Enterprise GPUA100 80GB SXM480$12,000
Enterprise GPUIntel Gaudi 3128$15,500
Enterprise GPUTesla V100 32GB32$19,000
Enterprise GPUA4048$19,000
Enterprise GPUH100 80GB PCIe80$25,000
Enterprise GPUH100 80GB SXM580$30,000
Enterprise GPUH200 141GB SXM141$35,000
Enterprise GPU-serverH100 NVL (2×H100 PCIe)188$65,000
Enterprise GPU-serverHGX H100 (8×H100 SXM5)640$350,000
Enterprise GPU-serverDGX B200 (8×B200)1440$400,000
Enterprise GPU-serverGB200 NVL72 (72×GB200)13500$3,000,000

Övergripande Investeringsanalys

Analysera kostnaden för att köra en specifik AI-modell genom att köpa eller hyra hårdvara.

1. Definiera din AI-modell

2. Resultat & Kostnadsuppskattning

Beräknat VRAM-behov

3.3 GB

Uppskattad Inköpskostnad

$110

Baserat på den generella pristrenden för att köpa GPU:er med motsvarande VRAM.
Uppskattad Hyreskostnad (per timme)

$0.017 / timme

Baserat på den generella pristrenden för att hyra GPU-servrar med motsvarande VRAM.

Break-Even Punkt

Det blir mer kostnadseffektivt att KÖPA hårdvaran istället för att hyra den 24/7 efter cirka 276 dagar (0.76 år).

3. Analys: Budget → Prestanda

Alternativ: Köp

Uppskattad VRAM för din budget

48.2 GB

Detta motsvarar en modellstorlek på:

PrecisionParametrar (Miljarder)
INT4 (0.5 byte)103.6B
INT8 (1 byte)51.8B
FP16/BF16 (2 byte)25.9B
FP32 (4 byte)12.9B

Alternativ: Hyr 24/7

Uppskattad VRAM för din budget

40.9 GB

Detta motsvarar en modellstorlek på:

PrecisionParametrar (Miljarder)
INT4 (0.5 byte)87.8B
INT8 (1 byte)43.9B
FP16/BF16 (2 byte)22.0B
FP32 (4 byte)11.0B

Den Gyllene Analysen: Lönsamhetskartor

Dessa värmekartor ger en översikt över hur ''investeringsbudget'' och ''daglig token-volym'' samverkar för att påverka brytpunktstiden. Använd kontrollerna för att definiera ditt analysområde och identifiera de mest lönsamma zonerna för din specifika situation.

Lönsamhetskarta för: Köp vs. API