Models / Best of /cheapest
Updated August 2026 · ranked from live data
Cheapest hosted models
Ranked by the lowest tracked output price per million tokens across all providers. Prices sync daily.Right now the top pick is Llama 3.2 1B Instruct ($0.020/M out).Every row links to full specs, hardware verdicts and provider pricing.
1Llama 3.2 1B InstructMeta$0.020/M out1.2B$0.020 / $0.020per 1M · via Novita AI2Ling-2.6-flashinclusionAI$0.030/M out107B$0.010 / $0.030per 1M · via OpenRouter3Mistral NemoMistral AI$0.030/M out12.2B$0.019 / $0.030per 1M · via OpenRouter4Llama 3 8B LunarisSao10K$0.050/M out8B$0.040 / $0.050per 1M · via OpenRouter5MythoMax 13BGryphe$0.060/M out13B$0.060 / $0.060per 1M · via OpenRouter6Mistral Small 3Mistral AI$0.080/M out23.6B$0.050 / $0.080per 1M · via OpenRouter7Llama 3.1 8B InstructMeta$0.080/M out8B$0.050 / $0.080per 1M · via OpenRouter8Nex-N2-MiniNex AGI$0.10/M out35.1B$0.025 / $0.10per 1M · via OpenRouter9Granite 4.1 8BIBM$0.10/M out8.8B$0.050 / $0.10per 1M · via OpenRouter10Reka EdgeReka AI$0.10/M out7.1B$0.10 / $0.10per 1M · via OpenRouter
