AIKraft Agents
ВойтиТарифы и покупка
AIKraft / Модели / Самые дешёвые AI-модели

Самые дешёвые AI-модели

Цена за миллион токенов различается в сотни раз, а разница в качестве на типовых задачах — куда меньше. Для классификации писем, извлечения полей из документов и черновиков дешёвая модель часто закрывает задачу полностью.

120 моделей в подборке · данные сверяются автоматически · весь каталог моделей

МодельКомпанияИнтеллектЦена / 1MКонтекстТокенов/с
Gemma 4 E4B (Non-reasoning) Google — $0.04 128K 46
Gemma 4 E4B (Reasoning) Google — $0.04 128K 48
Llama 3.1 Instruct 8B Meta — $0.04 128K 0
Mistral Small 3 Mistral — $0.06 32K 0
Qwen3.5 4B (Non-reasoning) Alibaba — $0.06 256K 37
Qwen3.5 4B (Reasoning) Alibaba — $0.06 256K 50
Nova Micro Amazon — $0.06 130K 246
Llama 3 Instruct 8B Meta — $0.07 8K 0
NVIDIA Nemotron Nano 9B V2 (Reasoning) NVIDIA — $0.07 128K 172
Sarvam 105B (high) Sarvam — $0.07 128K 131
Agnes 3.0 Flash — 35.5 $0.07 1M 0
NVIDIA Nemotron Nano 9B V2 (Non-reasoning) NVIDIA — $0.09 128K 172
NVIDIA Nemotron 3 Nano 30B A3B (Reasoning) NVIDIA — $0.09 1M 245
NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning) NVIDIA — $0.09 1M 267
gpt-oss-20b (high) OpenAI — $0.09 128K 138
Mistral Small 3.1 Mistral — $0.09 128K 0
Nemotron 3.5 Lightning NVIDIA 5.5 $0.10 1M 271
Hy3-preview (Non-reasoning) Tencent — $0.10 256K 0
Ministral 3 3B Mistral — $0.10 256K 181
Hy3-preview (Reasoning) Tencent — $0.10 256K 0
Nova Lite Amazon — $0.10 300K 0
Mistral Small 3.2 Mistral — $0.11 128K 0
Granite 4.2 8B IBM 5.1 $0.11 131K 75
Ling 3.0 Flash InclusionAI — $0.11 256K 0
DeepSeek V4 Flash (Non-reasoning) DeepSeek — $0.12 1M 0
Reka Flash 3 Reka AI — $0.13 128K 90
gpt-oss-20b (low) OpenAI — $0.13 128K 191
Phi-4 Mini Instruct Microsoft — $0.13 128K 46
GPT-5 nano (minimal) OpenAI — $0.14 400K 0
GPT-5 nano (medium) OpenAI — $0.14 400K 0
GPT-5 nano (high) OpenAI — $0.14 400K 0
Phi-4 Multimodal Instruct Microsoft — $0.14 128K 17
DeepSeek V4 Flash (Reasoning, High Effort) DeepSeek — $0.14 1M 0
Qwen3 30B A3B 2507 Instruct Alibaba — $0.15 256K 0
Qwen3 14B (Non-reasoning) Alibaba — $0.15 32K 0
Solar Mini Upstage — $0.15 4K 0
Gemma 4 12B (Reasoning) Google — $0.15 256K 139
MiMo-V2-Flash (Reasoning) Xiaomi — $0.15 256K 0
Qwen3 14B (Reasoning) Alibaba — $0.15 32K 0
Step 3.5 Flash StepFun — $0.15 256K 0
Ministral 3 8B Mistral — $0.15 256K 74
Gemma 4 12B (Non-reasoning) Google — $0.15 256K 140
Step 3.5 Flash 2603 StepFun — $0.15 256K 0
Qwen3.5 9B (Reasoning) Alibaba — $0.15 256K 83
GLM-4.7-Flash (Reasoning) Z AI — $0.15 200K 0
GLM-4.7-Flash (Non-reasoning) Z AI — $0.15 200K 0
Gemma 4 26B A4B (Reasoning) Google — $0.16 256K 97
Mistral 7B Instruct Mistral — $0.16 8K 0
Quasar 438B (max, based on GLM-5.2) Multiverse Computing — $0.17 1M 140
MiMo-V2.5 Xiaomi — $0.17 1M 0
Gemini 2.5 Flash-Lite (Non-reasoning) Google — $0.17 1M 0
DeepSeek V4 Flash 0731 (Reasoning, Max Effort) DeepSeek — $0.17 1M 0
Gemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning) Google — $0.17 1M 0
Gemini 2.5 Flash-Lite (Reasoning) Google — $0.17 1M 0
MiMo-V2.6-Flash Xiaomi 26.7 $0.17 1M 30
GPT-4.1 nano OpenAI — $0.17 1M 0
Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning) Google — $0.17 1M 0
Gemma 4 26B A4B (Non-reasoning) Google — $0.19 256K 107
Qwen3.5 9B (Non-reasoning) Alibaba — $0.19 256K 125
GPT-6 Luna (low) OpenAI 9.7 $0.20 1M 112
GPT-6 Luna (high) OpenAI 21.4 $0.20 1M 119
Ministral 3 14B Mistral — $0.20 256K 59
GPT-6 Luna (Non-reasoning) OpenAI 9.4 $0.20 1M 107
GPT-6 Luna (medium) OpenAI 18.5 $0.20 1M 0
Qwen3 8B (Non-reasoning) Alibaba — $0.20 32K 0
GPT-6 Luna (max) OpenAI 27.9 $0.20 1M 126
GPT-6 Luna (xhigh) OpenAI 23.5 $0.20 1M 117
Qwen3 30B A3B (Non-reasoning) Alibaba — $0.21 32K 0
Phi-4 Microsoft — $0.22 16K 44
Gemma 4 31B (Non-reasoning) Google — $0.22 256K 66
Qwen3.8-Flash-Next Alibaba — $0.23 256K 50
GLM 5.3 Flash Z AI 27.8 $0.24 1M 63
Qwen3 30B A3B (Reasoning) Alibaba — $0.24 32K 0
Hy3 Tencent — $0.24 256K 92
gpt-oss-120b (low) OpenAI — $0.24 128K 250
Ling-flash-2.0 InclusionAI — $0.25 128K 0
Jamba 1.5 Mini AI21 Labs — $0.25 256K 0
Solar Pro 3 Upstage — $0.26 128K 0
gpt-oss-120b (high) OpenAI — $0.26 128K 206
GPT-4o mini OpenAI — $0.26 128K 0
Mistral Small 4 (Reasoning) Mistral 4.3 $0.26 256K 163
Qwen3 32B (Non-reasoning) Alibaba — $0.26 32K 0
Mistral Small 4 (Non-reasoning) Mistral — $0.26 256K 155
Grok 4 Fast (Reasoning) SpaceXAI — $0.28 2M 0
Grok 4 Fast (Non-reasoning) SpaceXAI — $0.28 2M 0
Qwen3.5 Omni Flash Alibaba — $0.28 256K 231
Llama 4 Scout Meta — $0.30 10M 94
NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning) NVIDIA — $0.30 128K 206
Seed-OSS-36B-Instruct ByteDance Seed — $0.30 512K 0
DeepSeek V3.2 Exp (Reasoning) DeepSeek — $0.32 128K 0
DeepSeek V3.2 (Non-reasoning) DeepSeek — $0.32 128K 0
DeepSeek V3.2 Exp (Non-reasoning) DeepSeek — $0.32 128K 0
Qwen3 32B (Reasoning) Alibaba — $0.32 32K 0
DeepSeek V3.2 (Reasoning) DeepSeek — $0.32 128K 0
Celeris-1 — — $0.33 131K 1635
Qwen3 Coder 30B A3B Instruct Alibaba — $0.33 256K 0
Mercury 2.5 Inception — $0.34 260K 978
Llama 3.2 Instruct 11B (Vision) Meta — $0.34 128K 27
Qwen3 Next 80B A3B Instruct Alibaba — $0.35 256K 168
Reka Flash (Sep '24) Reka AI — $0.35 128K 0
Grok 3 mini Reasoning (high) SpaceXAI — $0.35 1M 0
GLM-4.5-Air Z AI — $0.37 128K 0
Mercury 2 Inception — $0.38 128K 0
Qwen3 235B A22B 2507 Instruct Alibaba — $0.39 256K 0
DeepSeek V3 0324 DeepSeek — $0.41 128K 0
Qwen3 VL 30B A3B Instruct Alibaba — $0.41 256K 0
Llama 4 Maverick Meta — $0.42 1M 60
Qwen3 Omni 30B A3B (Reasoning) Alibaba — $0.43 64K 111
Qwen3 Omni 30B A3B Instruct Alibaba — $0.43 64K 108
Step 3.7 Flash StepFun — $0.44 256K 0
GLM-4.6V (Non-reasoning) Z AI — $0.45 128K 0
GPT-5.6 Luna (high) OpenAI — $0.45 1M 0
GPT-5.6 Luna (max) OpenAI 23.8 $0.45 1M 0
Nemotron 3 Super 120B A12B (Reasoning) NVIDIA — $0.45 1M 0
GPT-5.6 Luna (medium) OpenAI — $0.45 1M 0
GLM-4.6V (Reasoning) Z AI — $0.45 128K 0
GPT-5.6 Luna (xhigh) OpenAI — $0.45 1M 0
GPT-5.6 Luna (low) OpenAI — $0.45 1M 0
GPT-5.6 Luna (Non-reasoning) OpenAI — $0.45 1M 0
DeepSeek V3.1 Terminus (Non-reasoning) DeepSeek — $0.45 128K 0

Другие подборки

AI-модели с рассуждениями (reasoning)Открытые AI-модели (open weights)AI-модели, доступные из РоссииСамые быстрые AI-моделиМодели с длинным контекстомМультимодальные AI-моделиРоссийские AI-моделиРейтингиВесь каталог

Компании в этой подборке

GoogleMetaMistralAlibabaAmazonNVIDIASarvamOpenAITencentIBMInclusionAIDeepSeekReka AIMicrosoft

Нужно выбрать модель под конкретную задачу компании и посчитать стоимость в рублях? Разбираем на бесплатном 30-минутном ИИ-аудите: какие процессы окупятся первыми и что это будет стоить.

Получить ИИ-аудит