SaaS Pricing Tracker

AI infrastructure and model inference pricing

The layer between a foundation model and an application: vector databases for retrieval, agent frameworks for orchestration, and inference providers that serve open-weight models by the token. Pricing here is usage-based almost without exception, metered by tokens, queries or storage rather than sold as a flat monthly seat.

AI infrastructure and model inference compared

Current list prices as last read from each pricing page.

ProductStarts atTop listed planPriced byFree planHistory
Arize AI$50 per monthAX Pro, USD per monthsingle paid planFlat per accountYesread Sep 2026
AssemblyAI$0.15 per hour of audioUniversal-3.5, USD per hour of audio$0.21 per hour of audioUniversal-3.5 ProunknownNoread Sep 2026
Braintrust$249 per month, $100 credits included, then usagePro, USD per month, platform feesingle paid planFlat per accountYesread Sep 2026
Cohere$2500 per month per instance, or $4.00 per hourEmbed 4 Small, USD per month per instance$3250 per month per instance, or $5.00 per hourRerank 3.5 MediumFlat per accountNoread Sep 2026
Deepgram$0.0057 per minute of audioFlux Multilingual, USD per minute of audio$0.0065 per minute of audioNova-3unknownNoread Sep 2026
Fireworks AI$0.008 per million tokensServerless, up to 150M params, USD per million tokens$0.1 per million tokensQwen3 8BunknownNoread Sep 2026
Galileo$100 per month, billed yearlyPro, USD per month, billed yearlysingle paid planFlat per accountYesread Sep 2026
Groq$0.05 per million input tokens, $0.08 per million outputLlama 3.1 8B Instant, USD per million input tokens$0.59 per million input tokens, $0.79 per million outputLlama 3.3 70B VersatileunknownNoread Sep 2026
Helicone$79 per monthPro, USD per month$799 per monthTeamFlat per accountYesread Sep 2026
Humanloopnot extractedsingle paid planContact salesNoread Sep 2026
LangChain$39 per seat per month, then pay as you goPlus, USD per seat per month, before usagesingle paid planPer seatYesread Sep 2026
Langfuse$29 per monthCore, USD per month$2499 per monthEnterpriseFlat per accountNoread Sep 2026
LangWatch€29 per core seat per month, 200K events includedGrowth, EUR per core seat per monthsingle paid planPer seatYesread Sep 2026
Modal$0.000164 per second of computeNvidia T4, USD per second of compute$0.001097 per second of computeNvidia H100 SXM5unknownNoread Sep 2026
Pinecone$20 per month, flat, solo developersBuilder, USD per month$500 per month, minimum usageEnterpriseFlat per accountYesread Sep 2026
Portkey$49 per month, 100K logs includedProduction, USD per monthsingle paid planFlat per accountYesread Sep 2026
Replicate$0.025 per output imageFLUX.1 schnell, USD per output image$0.04 per output imageFLUX.1 devunknownNoread Sep 2026
Together AI$0.35 per million input tokens, $1.50 per million outputMuse Glimmer 30B, USD per million input tokens$3 per million input tokens, $15.00 per million outputKimi K3unknownNoread Sep 2026
Weaviate$45 per month, starts at, pay-as-you-goFlex, USD per month, starting at$400 per month, starts at, prepaid contractPremiumFlat per accountYesread Sep 2026

AI infrastructure and model inference pricing questions

What does a AI infrastructure platform cost?
Paid plans start at $50 per month for Arize AI (USD per month), $0.15 per hour of audio for AssemblyAI (USD per hour of audio), $249 per month, $100 credits included, then usage for Braintrust (USD per month, platform fee), $2500 per month per instance, or $4.00 per hour for Cohere (USD per month per instance), $0.0057 per minute of audio for Deepgram (USD per minute of audio), $0.008 per million tokens for Fireworks AI (USD per million tokens), $100 per month, billed yearly for Galileo (USD per month, billed yearly), $0.05 per million input tokens, $0.08 per million output for Groq (USD per million input tokens), $79 per month for Helicone (USD per month), $39 per seat per month, then pay as you go for LangChain (USD per seat per month, before usage), $29 per month for Langfuse (USD per month), €29 per core seat per month, 200K events included for LangWatch (EUR per core seat per month), $0.000164 per second of compute for Modal (USD per second of compute), $20 per month, flat, solo developers for Pinecone (USD per month), $49 per month, 100K logs included for Portkey (USD per month), $0.025 per output image for Replicate (USD per output image), $0.35 per million input tokens, $1.50 per million output for Together AI (USD per million input tokens), $45 per month, starts at, pay-as-you-go for Weaviate (USD per month, starting at). Prices are as shown on each pricing page when it was last read.
What is AI infrastructure platform pricing based on?
Arize AI and Braintrust and Cohere and Galileo and Helicone and Langfuse and Pinecone and Portkey and Weaviate: flat per account. Humanloop: contact sales. LangChain and LangWatch: per seat.
Which ai infrastructure and model inference have a free plan?
Arize AI, Braintrust, Galileo, Helicone, LangChain, LangWatch, Pinecone, Portkey, Weaviate list a free plan next to the paid tiers.

Other categories