LIVE
DEEPSEEK-V4-FL$0.20 26.3%
DEEPSEEK-V4-PR$0.87 4.6%
QWEN3.6-FLASH$1.13 24.9%
NEMOTRON-3-SUP$0.45 23.9%
LLAMA-4-MAVERI$0.60 11.9%
LLAMA-4-SCOUT$0.30 5.8%
GEMINI-3.1-FLA$1.50 10.0%
GEMINI-2.5-FLA$0.40 0.8%
MINIMAX-01$1.10 8.0%
MIMO-V2.5$0.28 5.8%
MIMO-V2.5-PRO$0.87 2.6%
MINIMAX-M3$1.20 6.1%
QWEN3.5-PLUS-2$1.80 6.0%
NOVA-2-LITE-V1$2.50 31.4%
GEMINI-2.5-FLA$2.50 9.5%
GROK-4.3$2.50 8.7%
QWEN3.6-PLUS$1.95 31.6%
NEMOTRON-3-ULT$2.50 25.9%
QWEN3.7-PLUS$1.60 21.4%
MINIMAX-M1$2.20 11.9%
PALMYRA-X5$6.00 26.9%
QWEN3.7-MAX$3.75 3.1%
GEMINI-3.5-FLA$9.00 8.3%
GEMINI-2.5-PRO$10.00 7.2%
GPT-5.4-NANO$1.25 10.0%
NOVA-LITE-V1$0.24 28.9%
KIMI-K2.5$1.90 9.8%
MINISTRAL-14B-$0.20 17.7%
DEEPSEEK-V4-FL$0.20 26.3%
DEEPSEEK-V4-PR$0.87 4.6%
QWEN3.6-FLASH$1.13 24.9%
NEMOTRON-3-SUP$0.45 23.9%
LLAMA-4-MAVERI$0.60 11.9%
LLAMA-4-SCOUT$0.30 5.8%
GEMINI-3.1-FLA$1.50 10.0%
GEMINI-2.5-FLA$0.40 0.8%
MINIMAX-01$1.10 8.0%
MIMO-V2.5$0.28 5.8%
MIMO-V2.5-PRO$0.87 2.6%
MINIMAX-M3$1.20 6.1%
QWEN3.5-PLUS-2$1.80 6.0%
NOVA-2-LITE-V1$2.50 31.4%
GEMINI-2.5-FLA$2.50 9.5%
GROK-4.3$2.50 8.7%
QWEN3.6-PLUS$1.95 31.6%
NEMOTRON-3-ULT$2.50 25.9%
QWEN3.7-PLUS$1.60 21.4%
MINIMAX-M1$2.20 11.9%
PALMYRA-X5$6.00 26.9%
QWEN3.7-MAX$3.75 3.1%
GEMINI-3.5-FLA$9.00 8.3%
GEMINI-2.5-PRO$10.00 7.2%
GPT-5.4-NANO$1.25 10.0%
NOVA-LITE-V1$0.24 28.9%
KIMI-K2.5$1.90 9.8%
MINISTRAL-14B-$0.20 17.7%
HotON Insights

The Price Gap: US vs China AI Models

Chinese model providers now sit well below US flagships on price. We measured the gap across 120 live models (US 61 / China 51).

US avg blended
$8.30/1M
China avg blended
$1.04/1M
US / China
8.0×
Cheapest (China)
$0.01/1M

A 8.0× gap, in real prices

On a typical 3:1 output-to-input blend, US models average about $8.30 per 1M tokens versus roughly $1.04 for Chinese models — about 8.0× apart. These figures come from OpenRouter's live pricing, updated daily; they are not estimates.

Why it's happening

Chinese providers (DeepSeek, Alibaba's Qwen, Zhipu, Kimi, MiniMax and others) lean into open weights and aggressive pricing, and combined with fierce domestic competition and a focus on inference efficiency they push prices very low. US flagships price more for capability and ecosystem. For the same class of task, the model you pick can change cost by an order of magnitude.

What it means for builders

If your workload is price-sensitive, Chinese models (input as low as $0.01/1M) are often the more economical starting point — but cheaper isn't automatically better. Weigh it against your quality bar, latency and compliance needs. In the price-vs-efficiency map below, up-and-to-the-left is better value; color marks the region.

Market Map

Price vs Efficiency

USChinaEU

Every tracked model plotted by input price (log scale) and composite efficiency. Toward the top-left means better value per dollar.

$0.1$1$10$10060708090100Input price ($ / 1M tokens, log scale)EfficiencyLlama 4 Scout · $0.08/1M · eff 96GPT-5.5 Pro · $30.00/1M · eff 82GPT-5.4 Pro · $30.00/1M · eff 82GPT-5.5 · $5.00/1M · eff 82DeepSeek V4 Flash · $0.10/1M · eff 96DeepSeek V4 Pro · $0.44/1M · eff 96Llama 4 Maverick · $0.15/1M · eff 96Gemini 3.1 Flash Lite · $0.25/1M · eff 96Gemini 2.5 Flash Lite · $0.10/1M · eff 96MiMo-V2.5 · $0.14/1M · eff 96MiMo-V2.5-Pro · $0.44/1M · eff 96MiniMax M3 · $0.30/1M · eff 96Gemini 2.5 Flash · $0.30/1M · eff 95Gemini 3.5 Flash · $1.50/1M · eff 92Gemini 2.5 Pro · $1.25/1M · eff 92Palmyra X5 · $0.60/1M · eff 94Qwen3.6 Flash · $0.19/1M · eff 96Nemotron 3 Super · $0.09/1M · eff 96MiniMax-01 · $0.20/1M · eff 96Qwen3.5 Plus 2026-04-20 · $0.30/1M · eff 95Nova 2 Lite · $0.30/1M · eff 95Grok 4.3 · $1.25/1M · eff 95Qwen3.6 Plus · $0.33/1M · eff 95Nemotron 3 Ultra · $0.50/1M · eff 95Qwen3.7 Plus · $0.40/1M · eff 95MiniMax M1 · $0.40/1M · eff 95Qwen3.7 Max · $1.25/1M · eff 94Nova Premier 1.0 · $2.50/1M · eff 90Claude Sonnet 4.6 · $3.00/1M · eff 89Claude Opus 4.8 · $5.00/1M · eff 84GPT-5.4 Nano · $0.20/1M · eff 91GPT-5.4 Mini · $0.75/1M · eff 89GPT-5.3-Codex · $1.75/1M · eff 85GPT-5.2-Codex · $1.75/1M · eff 85Nova Lite 1.0 · $0.06/1M · eff 91Nova Pro 1.0 · $0.80/1M · eff 89Kimi K2.5 · $0.40/1M · eff 90Ministral 3 14B 2512 · $0.20/1M · eff 90Ring-2.6-1T · $0.08/1M · eff 90Mistral Small 4 · $0.15/1M · eff 90MiMo-V2-Flash · $0.10/1M · eff 90Seed-2.0-Lite · $0.25/1M · eff 90Qwen3.5-9B · $0.04/1M · eff 90Ling-2.6-flash · $0.01/1M · eff 90Seed-2.0-Mini · $0.10/1M · eff 90Seed 1.6 · $0.25/1M · eff 90Qwen3.6 35B A3B · $0.14/1M · eff 90Step 3.5 Flash · $0.09/1M · eff 90Nemotron 3 Nano 30B A3B · $0.05/1M · eff 90Seed 1.6 Flash · $0.08/1M · eff 90Ling-2.6-1T · $0.08/1M · eff 90Mistral Large 3 2512 · $0.50/1M · eff 90Gemma 4 26B A4B · $0.06/1M · eff 90Gemma 4 31B · $0.12/1M · eff 90Qwen3.6 27B · $0.29/1M · eff 89Kimi K2.6 · $0.68/1M · eff 89Devstral 2 2512 · $0.40/1M · eff 89Kimi K2 0905 · $0.60/1M · eff 89Mistral Medium 3.5 · $1.50/1M · eff 87Codestral 2508 · $0.30/1M · eff 90Step 3.7 Flash · $0.20/1M · eff 90Jamba Large 1.7 · $2.00/1M · eff 87Command A · $2.50/1M · eff 86MiniMax M2 · $0.26/1M · eff 90MiniMax M2.1 · $0.29/1M · eff 90MiniMax M2.7 · $0.28/1M · eff 89MiniMax M2.5 · $0.15/1M · eff 89GLM 4.7 Flash · $0.06/1M · eff 90GLM 5.1 · $0.98/1M · eff 88GLM 5 Turbo · $1.20/1M · eff 88GLM 5V Turbo · $1.20/1M · eff 88Claude 3.5 Haiku · $0.80/1M · eff 88Claude Haiku 4.5 · $1.00/1M · eff 88Sonar Pro · $3.00/1M · eff 83Sonar Pro Search · $3.00/1M · eff 83DeepSeek V3.1 · $0.21/1M · eff 89R1 0528 · $0.50/1M · eff 89DeepSeek V3.1 Terminus · $0.27/1M · eff 89Llama 3.3 70B Instruct · $0.10/1M · eff 89Granite 4.1 8B · $0.05/1M · eff 89GLM 4.6V · $0.30/1M · eff 89Trinity Mini · $0.05/1M · eff 89DeepSeek V3.2 · $0.23/1M · eff 89Phi 4 Mini Instruct · $0.08/1M · eff 89Llama 3.3 Nemotron Super 49B V1.5 · $0.10/1M · eff 89Hermes 4 70B · $0.13/1M · eff 89ERNIE 4.5 VL 28B A3B · $0.14/1M · eff 89GLM 4.5 Air · $0.13/1M · eff 89Hunyuan A13B Instruct · $0.14/1M · eff 89ERNIE 4.5 VL 424B A47B · $0.42/1M · eff 89Nemotron Nano 9B V2 · $0.04/1M · eff 89Spotlight · $0.18/1M · eff 89Virtuoso Large · $0.75/1M · eff 89Llama 3.2 3B Instruct · $0.05/1M · eff 89R1 Distill Llama 70B · $0.70/1M · eff 89Granite 4.0 Micro · $0.02/1M · eff 89Maestro Reasoning · $0.90/1M · eff 88LFM2-24B-A2B · $0.03/1M · eff 89Solar Pro 3 · $0.15/1M · eff 89GLM 4 32B · $0.10/1M · eff 89UI-TARS 7B · $0.10/1M · eff 89Mistral Small 3.2 24B · $0.08/1M · eff 89R1 Distill Qwen 32B · $0.29/1M · eff 89Command R7B (12-2024) · $0.04/1M · eff 89Nova Micro 1.0 · $0.04/1M · eff 89Sonar Reasoning Pro · $2.00/1M · eff 86Sonar Deep Research · $2.00/1M · eff 86GPT-5.3 Chat · $1.75/1M · eff 83Sonar · $1.00/1M · eff 89Olmo 3 32B Think · $0.15/1M · eff 89Reka Flash 3 · $0.10/1M · eff 89GLM 4.5V · $0.60/1M · eff 88MiniMax M2-her · $0.30/1M · eff 88Gemma 3n 4B · $0.06/1M · eff 89Coder Large · $0.50/1M · eff 88Voxtral Small 24B 2507 · $0.10/1M · eff 89Phi 4 · $0.07/1M · eff 88Reka Edge · $0.10/1M · eff 88Inflection 3 Pi · $2.50/1M · eff 84Inflection 3 Productivity · $2.50/1M · eff 84

Each dot is one model · color = region · click a dot to open it.

Pricing is real (via OpenRouter, updated daily). This is market analysis, not investment or procurement advice.