Intelligence des modèles d'IA

Tarifs · 2026-05-12

Cheapest LLM APIs

AI APIs ranked by input + output token cost.

À propos de cette liste

  • All text-in / text-out LLM APIs sorted from cheapest to most expensive by total per-token cost.
  • Models with $0 placeholder pricing (free promo tiers, GitHub Copilot rebroadcasts) are excluded — Unknown is not the same as free.
  • Use this list to find the lowest-cost option that still meets your context and capability requirements.
#ModèleÉditeurEntrée / 1MSortie / 1MTotalContexte
1BGE Reranker Basecloudflare-ai-gateway$0.003Unknown$0.003128K
2Voxtral Small 24B 2507Mistral$0.002$0.002$0.00532K
3Multi-QA-mpnet-base-dot-v1digitalocean$0.009Unknown$0.009512
4All-MiniLM-L6-v2digitalocean$0.009Unknown$0.009256
5Qwen3 Embedding 8BAlibaba (Qwen)$0.010Unknown$0.01033K
6Qwen3 Embedding 4BAlibaba (Qwen)$0.010Unknown$0.01033K
7Qwen3 Embedding 0.6BAlibaba (Qwen)$0.010Unknown$0.01033K
8BGE Reranker v2 M3digitalocean$0.010Unknown$0.0108K
9BGE M3cloudflare-ai-gateway$0.012Unknown$0.012128K
10PLaMo Embedding 1Bcloudflare-ai-gateway$0.019Unknown$0.019128K
11Llama 3.2 1B InstructMeta$0.010$0.010$0.02016K
12Llama Prompt Guard 2 22MMeta$0.010$0.010$0.020512
13Llama Prompt Guard 2 86MMeta$0.010$0.010$0.020512
14text-embedding-3-smallOpenAI$0.020Unknown$0.0208K
15BGE Small EN v1.5cloudflare-ai-gateway$0.020Unknown$0.020128K
16text-embedding-3-smallazure-cognitive-services$0.020Unknown$0.0208K
17E5 Large v2digitalocean$0.020Unknown$0.020512
18BGE M3digitalocean$0.020Unknown$0.0208K
19voyage-3.5-litevoyageai$0.020Unknown$0.0208K
20Titan Text Embeddings V2vercel$0.020Unknown$0.0208K
21text-embedding-3-smallazure$0.020Unknown$0.0208K
22dots.ocrchutes$0.010$0.011$0.021131K
23Llama 3.2 3b InstructMeta$0.010$0.014$0.024131K
24DistilBERT SST-2 INT8cloudflare-ai-gateway$0.026Unknown$0.026128K
25Gemma 3 4BGoogle$0.010$0.027$0.03733K
26PaddleOCR-VLnovita-ai$0.020$0.020$0.04016K
27Meta-Llama-3.1-8B-InstructMeta$0.020$0.030$0.050128K
28Llama 3.1 8B TurboMeta$0.020$0.030$0.050131K
29Mistral Nemo Instruct 2407Mistral$0.020$0.040$0.060128K
30Gemma 3n 4BGoogle$0.020$0.040$0.0608K
31voyage-3.5voyageai$0.060Unknown$0.0608K
32BGE Base EN v1.5cloudflare-ai-gateway$0.067Unknown$0.067128K
33Hermes 4 14Bchutes$0.014$0.054$0.06841K
34Meta-Llama-3-8B-InstructMeta$0.030$0.040$0.0708K
35Llama Guard 3 8BMeta$0.020$0.060$0.0808K
36Ministral 3Bazure-cognitive-services$0.040$0.040$0.080128K
37Ministral 3B (latest)Mistral$0.040$0.040$0.080128K
38Llama 3.1 8B InstructMeta$0.030$0.050$0.080131K
39Ministral 3Bazure$0.040$0.040$0.080128K
40Sao10K: Llama 3 8B LunarisMeta$0.040$0.050$0.0908K
41GTE Large (v1.5)digitalocean$0.090Unknown$0.0908K
42Llama-3.2-11B-Vision-InstructMeta$0.049$0.049$0.098128K
43text-embedding-ada-002OpenAI$0.100Unknown$0.1008K
44Mistral EmbedMistral$0.100Unknown$0.1008K
45Sao10k L3 8B Lunaris novita-ai$0.050$0.050$0.1008K
46L3 8B Stheno V3.2novita-ai$0.050$0.050$0.1008K
47text-embedding-ada-002azure-cognitive-services$0.100Unknown$0.1008K
48text-embedding-ada-002azure$0.100Unknown$0.1008K
49MythoMax 13Bkilo$0.060$0.060$0.1204K
50voyage-finance-2voyageai$0.120Unknown$0.1208K
51voyage-code-2voyageai$0.120Unknown$0.1208K
52voyage-law-2voyageai$0.120Unknown$0.1208K
53IBM: Granite 4.0 Microkilo$0.017$0.110$0.127131K
54IBM Granite 4.0 H Microcloudflare-ai-gateway$0.017$0.110$0.127128K
55Gemma 3 12BGoogle$0.030$0.100$0.13033K
56Llama 3.1 8B InstantMeta$0.050$0.080$0.130131K
57text-embedding-3-largeOpenAI$0.130Unknown$0.1308K
58text-embedding-3-largeazure-cognitive-services$0.130Unknown$0.1308K
59Llama 3 8BMeta$0.050$0.080$0.1308K
60text-embedding-3-largeazure$0.130Unknown$0.1308K
61Gemma 3 27BGoogle$0.027$0.109$0.136131K
62Qwen2.5-Coder 32B InstructAlibaba (Qwen)$0.027$0.109$0.136131K
63DeepSeek R1 Distill Llama 70BDeepSeek$0.027$0.109$0.1368K
64baichuan-m2-32bnovita-ai$0.070$0.070$0.140131K
65Model Routerazure-cognitive-services$0.140Unknown$0.140128K
66Model Routerazure$0.140Unknown$0.140128K
67Gemini Embedding 001Google$0.150Unknown$0.1502K
68LiquidAI: LFM2-24B-A2Bkilo$0.030$0.120$0.15033K
69IBM: Granite 4.1 8Bkilo$0.050$0.100$0.150131K
70DeepSeek R1 Distill Llama 70BMeta$0.030$0.130$0.160131K
71GPT OSS 20BOpenAI$0.030$0.140$0.170131K
72GLM Z1 9B 0414Z.AI / Zhipu$0.086$0.086$0.17232K
73GLM 4 9B 0414Z.AI / Zhipu$0.086$0.086$0.17232K
74Amazon: Nova Micro 1.0kilo$0.035$0.140$0.175128K
75Nova Microamazon-bedrock$0.035$0.140$0.175128K
76Nova Microvercel$0.035$0.140$0.175128K
77Amazon Nova Micro 1.0nano-gpt$0.036$0.139$0.175128K
78Phi 4 Multimodalnano-gpt$0.070$0.110$0.180128K
79Manta Mini 1.0nano-gpt$0.020$0.160$0.1808K
80Manta Flash 1.0nano-gpt$0.020$0.160$0.18016K

Top 80 sur 924 affichés. Voir le reste dans le répertoire complet.

Frequently asked questions

What is the cheapest LLM API right now?

BGE Reranker Base is the lowest-priced AI APIs on this list, at $0.003 per 1M input tokens and Unknown per 1M output tokens. The total column above sums input + output per 1M tokens for direct comparison.

Why are prices shown per 1M tokens?

Almost every commercial LLM provider publishes rates per million tokens. Per-1K-tokens or per-token rates are easy to misread (off by 1000×). Standardising on $/1M lets you compare vendors directly.

Are these prices including or excluding tax?

All prices are the provider's headline list rate in USD, before any volume discounts, prepaid credits, batch-API discounts, prompt-caching discounts or jurisdictional sales tax. Always verify with the provider before committing to a budget.

How often do these prices change?

Models.dev pushes price changes within hours of a vendor announcement, and our pipeline re-syncs daily. We also write each change to /changelog so you can audit historical pricing over time.

Why are some models showing 'Unknown' instead of a price?

We deliberately do not coerce missing data to $0. 'Unknown' means the provider does not publish a public rate (often models behind enterprise sales or invite-only access). Treating Unknown as free would push paid-but-unpriced models to the top of every cheap list — broken UX and broken SEO.

Dernière mise à jour :

Prices in USD per 1M tokens. Unknown means the provider does not publish per-token pricing.

Data is sourced from models.dev and normalized for comparison. Prices and capabilities may change. Always verify critical production decisions with the provider's official documentation.