Advertise
The race

🇺🇸 US vs 🇨🇳 China vs 🇪🇺 Europe

Top US pick: GPT-6 Astra. Top Chinese pick: Kimi K3. In our editorial view, US labs hold the frontier tier while Chinese labs lead on price and open weights. Europe coverage is still limited.

Models tracked

🇺🇸 US
14
🇨🇳 CN
8
🇪🇺 EU
1

Frontier-tier models

🇺🇸 US
4
🇨🇳 CN
1
🇪🇺 EU
0

Open-weight models

🇺🇸 US
1
🇨🇳 CN
6
🇪🇺 EU
1

Avg output price /1M

🇺🇸 US
$12.88
🇨🇳 CN
$4.09
🇪🇺 EU
$0.60
USChinaEurope
BasicLiteStrongProFrontier$0.1$0.3$1$3$10$30Claude Haiku 5.5: $0.1/$0.5 per 1MClaude Haiku 5.5GPT-6 Luna: $0.1/$0.5 per 1MGPT-6 LunaQwen3.8 Flash: $0.15/$0.47 per 1MQwen3.8 FlashGLM-5.3-Flash: $0.15/$0.5 per 1MGLM-5.3-FlashMistral Small 4: $0.15/$0.6 per 1MMistral Small 4gpt-oss-120B: $0.15/$0.6 per 1Mgpt-oss-120BDeepSeek V4.1 Flash: $0.3/$1.2 per 1MDeepSeek V4.1 FlashMiniMax-M3: $0.3/$1.2 per 1MMiniMax-M3Gemini 3.5 Flash-Lite: $0.3/$2.5 per 1MGemini 3.5 Flash-LiteGemini 3.8 Flash: $0.75/$3.75 per 1MGemini 3.8 FlashGrok 4.3: $1.25/$2.5 per 1MGrok 4.3DeepSeek V4 Pro: $1.32/$3.96 per 1MDeepSeek V4 ProGLM-5.3: $1.4/$4.4 per 1MGLM-5.3Grok 4.7: $2/$6 per 1MGrok 4.7Qwen3.8-Max: $2/$6 per 1MQwen3.8-MaxGemini 4 Argon: $2/$10 per 1MGemini 4 ArgonGPT-6.1 Sol: $2/$10 per 1MGPT-6.1 SolClaude Sonnet 5.5: $2/$10 per 1MClaude Sonnet 5.5GPT-5.3-Codex: $1.75/$14 per 1MGPT-5.3-CodexKimi K3: $3/$15 per 1MKimi K3Claude Opus 5.5: $4/$20 per 1MClaude Opus 5.5GPT-6 Astra: $10/$50 per 1MGPT-6 AstraClaude Fable 5.1: $10/$50 per 1MClaude Fable 5.1Blended price per 1M tokens (3:1 in:out, log scale) →

🇺🇸 United States

#ModelRegionTierPrice in / outContextStrengths
1GPT-6 AstraOpenAI🇺🇸 USFrontier$10 / $50UnknownTop-tier OpenAI pro modelsrc
2Claude Fable 5.1Anthropic🇺🇸 USFrontier$10 / $501MAnthropic's top modelsrc
3Claude Opus 5.5Anthropic🇺🇸 USFrontier$4 / $201MStrong coding & agentic worksrc
4Gemini 4 ArgonGoogle🇺🇸 USFrontier$2 / $101MGoogle flagship; intro price matches Sol/Sonnet until Nov 30
Introductory $2/$10 through 2026-11-30; then $4/$20 from Dec 1		src
5GPT-6.1 SolOpenAI🇺🇸 USPro$2 / $10UnknownFlagship value, cheap cached input ($0.10)src
6Claude Sonnet 5.5Anthropic🇺🇸 USPro$2 / $101MBalanced cost/qualitysrc
7Grok 4.7xAI🇺🇸 USPro$2 / $6500KxAI's most capable; low output pricesrc
8GPT-5.3-CodexOpenAI🇺🇸 USPro$1.75 / $14UnknownCoding-specializedsrc
9Grok 4.3xAI🇺🇸 USStrong$1.25 / $2.51MCheap, 1M contextsrc
10Gemini 3.8 FlashGoogle🇺🇸 USStrong$0.75 / $3.751MFast; promo price until 2026-12-31 (then $1.50/$7.50)src
11gpt-oss-120BOpenAI (via Together) · open🇺🇸 USStrong$0.15 / $0.6131KOpen-weights OpenAI modelsrc
12Claude Haiku 5.5Anthropic🇺🇸 USStrong$0.1 / $0.51MCheapest Claude (≤100K prompt; $0.50/$2.50 above)src
13GPT-6 LunaOpenAI🇺🇸 USStrong$0.1 / $0.51MLowest-price flagship family tiersrc
14Gemini 3.5 Flash-LiteGoogle🇺🇸 USLite$0.3 / $2.51MLow-latencysrc

🇨🇳 China

#ModelRegionTierPrice in / outContextStrengths
1Kimi K3Moonshot AI · open🇨🇳 CNFrontier$3 / $151MChina's top-priced reasoning model; open-weightsrc
2DeepSeek V4 ProDeepSeek · open🇨🇳 CNPro$1.32 / $3.961MLow price; 50% off-peak
Peak cache-miss rate; open-weights status: verify		src
3Qwen3.8-MaxAlibaba (Qwen)🇨🇳 CNPro$2 / $61MAlibaba flagship; thinking + non-thinking modes
International deployment price		src
4GLM-5.3Zhipu / Z.ai · open🇨🇳 CNPro$1.4 / $4.41MStrong value open-weight modelsrc
5DeepSeek V4.1 FlashDeepSeek · open🇨🇳 CNStrong$0.3 / $1.21MVery cheap; half off-peak
Open-weights status: verify		src
6MiniMax-M3MiniMax · open🇨🇳 CNStrong$0.3 / $1.21MVery cheap (permanent 50% off; 2x above 512K input)
Open-weights status: verify		src
7Qwen3.8 FlashAlibaba (Qwen)🇨🇳 CNLite$0.15 / $0.471MUltra-low pricesrc
8GLM-5.3-FlashZhipu / Z.ai · open🇨🇳 CNLite$0.15 / $0.51MCheapest GLM tier
Open-weights status: verify		src

🇪🇺 Europe

#ModelRegionTierPrice in / outContextStrengths
1Mistral Small 4Mistral · open🇪🇺 EULite$0.15 / $0.6262KCheap, efficient
Open-weights status: verify		src
AdvertisementTODO: AdSense slot
AdvertisementTODO: AdSense slot