Ranking · data through 2026-07-07

The cheapest LLM APIs, ranked.

Every provider’s lightest tier, ranked by list-price input rate, verified nightly against official pricing pages. As of 2026-07-07, the cheapest tracked rate is Llama 3.1 8B on DeepInfra at $0.02 per 1M input tokens.

#ProviderLightest tierInput / 1MOutput / 1MContextThe Hobbit*
1DeepInfraLlama 3.1 8B$0.02$0.05128K$0.0064
2CohereCommand R7B$0.0375$0.15128K$0.0191
3GroqLlama 3.1 8B$0.05$0.08128K$0.0102
4Together AIGPT-OSS 20B$0.05$0.2128K$0.0254
5Fireworks AIGPT-OSS 20B$0.07$0.3128K$0.0381
6OpenRouterLlama 3.3 70B$0.1$0.32128K$0.0407
7Google GeminiFlash-Lite 2.5$0.1$0.41M$0.0509
8Meta LlamaLlama 4 Scout$0.11$0.3410M$0.0432
9DeepSeekV4 Flash$0.14$0.281M$0.0356
10Mistral AIMistral Small 4$0.15$0.632K$0.0763
11Alibaba QwenQwen3.6-Flash$0.165$0.991M$0.1259
12Zhipu GLMGLM-4.5-Air$0.2$1.10200K$0.1399
13SambaNovaGPT-OSS 120B$0.22$0.59128K$0.075
14MiniMaxMiniMax M2$0.3$1.20200K$0.1526
15CerebrasGPT-OSS 120B$0.35$0.75128K$0.0954
16Moonshot KimiKimi K2.5$0.6$3.00256K$0.3814
17OpenAIGPT-5.4 mini$0.75$4.50128K$0.5721
18Perplexity†Sonar$1.00$1.00127K$0.1271
19xAI GrokGrok Build 0.1$1.00$2.00256K$0.2543
20AnthropicClaude Haiku 4.5$1.00$5.00200K$0.6357
21AWS BedrockClaude Haiku 4.5 (Bedrock)$1.00$5.00200K$0.6357
22Sakana AIFugu Ultra$5.00$30.001M$3.81

* Cost to generate The Hobbit (95,356 words ≈ 127,141 tokens) as output. † Perplexity bills web search separately per request; token prices exclude that fee.

How to read this table

About these numbers. Prices are checked nightly against each provider’s official pricing page and recorded in an append-only history (data through 2026-07-07). List prices only: promotions never enter the ranking. Corrections are published, not buried: see the methodology.

More pricing pages