AI Models

62 models from 14 providers · Prices shown per million tokens · 33 new models added

AMA Smart Routes · Built by Us

One endpoint that picks the right model

Use nexus/* routes and let our system automatically select the optimal model for each task — no model-switching overhead.

Smart Auto
nexus/auto

Smart routing — picks the best quality/cost balance among all live models for your request.

Smart Routing
nexus/cheapest-west

Routes to the cheapest capable Western provider model — balanced capability, speed, and cost.

Lowest Cost
nexus/cheapest

Always routes to the cheapest capable model for your task. Great for high-volume workloads.

CN Provider Optimized
nexus/cheapest-cn

Routes to the cheapest China-based provider models for eligible users in permitted regions.

Speed + Quality
nexus/best-reasoning

Optimizes for both response speed and output quality. Best for production chat apps.

Ultra-Fast
nexus/fastest

Always picks the fastest available model. Ideal for real-time streaming and low-latency apps.

Best for Code
nexus/best-coding

Routes to the strongest coding-capable model available. Best for refactors, reviews, and complex logic.

Long Context
nexus/long-context

Routes to the cheapest model with 500K+ context. Ideal for long documents and large codebases.

Web Search
nexus/web-search

Routes to Perplexity Sonar models with live web search built in. Best for research and current events.

Trending nowSeed3D 2.0 (CN)+70%Seedance 2.0 (CN)+62%Seed3D 2.0 (International)+60%

62 models shown

Provider
Use
Claude Sonnet 5NEWLatest
Latest Sonnet — claude-sonnet-latest routes here. Intro pricing through Aug 2026.
Anthropic
$1.90
$9.50
1M
96
80
+45%
GPT-5.6 SolNEWGPT-5.6 Flagship
OpenAI GPT-5.6 flagship — highest intelligence in the 5.6 family
OpenAI
$5.00
$30.00
1M
99
68
+22.4%
GPT-5.6 TerraNEWBalanced
GPT-5.6 balanced tier — strong quality at mid price
OpenAI
$2.50
$15.00
1M
96
76
+18.1%
GPT-5.6 LunaNEWFast GPT-5.6
GPT-5.6 fast/efficient tier for high-volume workloads
OpenAI
$1.00
$6.00
1M
92
88
+16.5%
Claude Opus 4.8NEWLive
Use claude-opus-4-8 or claude-opus-latest — platform routes to ArkLin automatically.
Anthropic
$4.75
$23.75
1M
98
66
+48.2%
Claude Opus 4.7BYOK / Soon
Anthropic catalog + BYOK direct API. Opus 4.7 on the US pool rolls out when enabled on the platform.
Anthropic
$4.75
$23.75
200K
97
68
+18%
GPT-5.5 ProMost Powerful
OpenAI's most powerful frontier model for the most complex tasks
OpenAI
$5.00
$30.00
256K
98
62
+4.2%
GPT-5.5Latest
OpenAI's flagship model for professional workloads (Apr 2026)
OpenAI
$2.50
$10.00
256K
95
74
+18.6%
Gemini 3.1 ProNEWLatest
Google's most capable model, 2M context, multimodal (Feb 2026)
Google
$2.00
$12.00
1M
94
72
+9.8%
Claude Fable 5NEWRestricted
Anthropic 第五代旗舰 — 上游限制许可,暂不可调用。请用 Sonnet 5 / Opus 4.8。
Anthropic
$10.00
$50.00
1M
99
64
0%
DeepSeek V4 ProNEW75% OFF
Advanced reasoning, 75% launch discount until May 31 — list price $1.74/$3.48
DeepSeek
$0.500
$1.00
1M
90
70
+44.5%
GPT-5.4Recommended
Current flagship GPT — powerful, 1M context, fast
OpenAI
$2.50
$10.00
128K
91
81
+3.1%
DeepSeek V4 FlashNEWCheapest
Fastest DeepSeek model — ultra-cheap for high-volume tasks
DeepSeek
$0.140
$0.280
1M
78
95
+19.3%
GPT-5.4 Mini
Cost-efficient GPT-5.4 class model, great for most tasks
OpenAI
$0.150
$0.600
128K
84
88
-1.2%
Hy3 PreviewNEWHunyuan 3
Tencent Hunyuan 3 (混元 3) — 295B MoE, agent-first, open-sourced Apr 2026
Tencent
$0.190
$0.620
262K
89
72
+52.1%
Seedance 2.0 (CN)NEWVideo · CN
Volcengine Ark CN — text/image/video-to-video, up to 15s, 720p/1080p. API: POST /v1/contents/generations/tasks
ByteDance
$0.0000
$0.0000
0K
94
55
+62%
Seedance 2.0 Fast (CN)NEWVideo · CN
Faster CN route for prompt iteration — 720p
ByteDance
$0.0000
$0.0000
0K
86
78
+48%
Seedance 2.0 (International)NEWVideo · Intl
BytePlus ModelArk international region — same API, base URL routed automatically
ByteDance
$0.0000
$0.0000
0K
94
55
+55%
Seedance 2.0 Fast (Intl)NEWVideo · Intl
International fast variant for draft renders
ByteDance
$0.0000
$0.0000
0K
86
78
+44%
Seed3D 2.0 (CN)NEW3D · CN
Volcengine Ark CN — image/text-to-3D asset generation. API: POST /v1/contents/generations/tasks
ByteDance
$0.0000
$0.0000
0K
92
50
+70%
Seed3D 1.0 (CN)NEW3D · CN
Seed3D 1.0 — image/text-to-3D (China region)
ByteDance
$0.0000
$0.0000
0K
85
55
+45%
Grok 4.3NEWLatest
xAI flagship — 1M context, configurable reasoning effort
xAI
$1.25
$2.50
1M
92
76
+17.9%
Claude Sonnet 4.6Most Popular
Best balance of Claude intelligence and speed — 1M context
Anthropic
$2.85
$14.25
1M
93
78
+11.4%
Gemini 3.1 Flash-LiteNEWBest Value
Best price-performance in Gemini family, retains free tier
Google
$0.250
$1.50
1M
80
94
+31.2%
Qwen3.7 MaxNEWLatest
Alibaba Qwen3.7 flagship — Bailian verified, 1M context
Alibaba
$0.800
$3.20
1M
91
76
+32%
GLM-5.1NEWLatest
Zhipu GLM-5.1 — Bailian/ArkLin verified live
Zhipu AI
$1.00
$3.20
200K
92
75
+30%
Kimi K2.7 CodeNEWLatest
Moonshot coding flagship — 256K context, best for agentic code
Moonshot
$0.950
$4.00
256K
90
78
+22.4%
Grok 4.1 FastLegacy
Legacy slug — xAI redirects to Grok 4.3 (none reasoning effort)
xAI
$1.25
$2.50
1M
79
93
+6.3%
Gemini 3.0 Flash
Fast next-gen model, proven stability, free tier available
Google
$0.575
$3.45
1M
79
91
-4.1%
DeepSeek R1
Reasoning alias → V4 Pro (replaces deepseek-reasoner)
DeepSeek
$0.550
$2.19
1M
87
66
-6.3%
Mistral Large
Top European AI — strong multilingual, GDPR-compliant
Mistral
$2.00
$6.00
131K
83
77
-1.8%
Sonar ProWeb Search
Real-time web search + reasoning — live internet access
Perplexity
$3.00
$15.00
127K
86
71
+7.2%
Qwen3.7 PlusNEWNew
Qwen3.7 plus — strong balance of quality and cost
Alibaba
$0.400
$1.60
1M
89
80
+26%
Qwen3.6 Plus
Qwen3.6 long-context — still widely used
Alibaba
$0.325
$1.95
1M
88
79
+12.4%
Kimi K2.6
Multimodal general model — vision, thinking, and agents
Moonshot
$0.550
$2.65
256K
86
75
+15.8%
Command R+
Enterprise RAG specialist — retrieval-augmented generation
Cohere
$2.50
$10.00
128K
82
73
-3.5%
GPT-5.4 NanoCheapest GPT
Ultra-fast, ultra-cheap — high-volume tasks and simple queries
OpenAI
$0.230
$1.44
1M
77
96
+2.4%
GPT-5
Base GPT-5 — excellent balance of capability and cost
OpenAI
$0.720
$5.75
1M
88
84
-3%
GPT-4o
Proven multimodal model — stable for existing integrations
OpenAI
$2.88
$11.50
128K
82
80
-8.4%
Sonar
Fast web-grounded answers — great for factual Q&A
Perplexity
$1.00
$1.00
127K
77
88
+3.1%
Claude Haiku 4.5Fast
Near-instant responses, high throughput, full 1M context
Anthropic
$0.950
$4.75
200K
83
93
+4.7%
Mistral Small
Lightweight, fast, EU-based for compliance-sensitive use
Mistral
$0.115
$0.345
33K
72
91
-0.4%
Llama 4 MaverickOpen Source
Open-source multimodal — 512K context, strong for its price
Meta
$0.220
$0.880
524K
81
82
+5.6%
Llama 4 Scout
Efficient open-source, long context, ultra-low cost
Meta
$0.092
$0.345
524K
74
89
+2.9%
GLM-5.2NEWAlias
GLM-5.2 catalog id — routes to live glm-5.1 until Bailian paid quota is enabled
Zhipu AI
$1.00
$4.00
1M
92
75
+28.5%
Qwen MaxBest Chinese
Alibaba's most powerful model — excels at Chinese tasks
Alibaba
$0.400
$1.20
33K
85
76
+1.5%
Qwen Plus
Balanced performance and cost in the Qwen family
Alibaba
$0.092
$0.299
131K
75
84
-2.1%
Qwen Turbo
Fastest, cheapest Qwen — high-volume Chinese language tasks
Alibaba
$0.050
$0.200
131K
68
97
+0.8%
GLM-4
Legacy alias — routes to GLM-5.1
Zhipu AI
$0.140
$0.140
128K
76
85
+1.1%
Claude Opus 4.8 RscNEWRsc · VIP
ArkLin VIP Rsc channel — Opus 4.8, dedicated high-priority route.
ArkLin
$2.00
$10.00
200K
98
78
+52%
Claude Opus 4.7 RscNEWRsc · VIP
ArkLin VIP Rsc channel — Opus 4.7.
ArkLin
$2.00
$10.00
200K
97
76
+44%
Claude Sonnet 5 RscNEWRsc · VIP
ArkLin VIP Rsc channel — Sonnet 5, speed + intelligence balance.
ArkLin
$0.800
$4.00
1M
96
84
+46%
Claude Sonnet 4.6 RscNEWRsc · VIP
ArkLin VIP Rsc channel — Sonnet 4.6.
ArkLin
$1.20
$6.00
200K
94
80
+36%
DeepSeek ChatLegacy
Legacy alias → V4 Flash (DeepSeek retires 2026-07-24)
DeepSeek
$0.140
$0.280
1M
78
95
+5%
Seed3D 2.0 (International)NEW3D · Intl
BytePlus ModelArk international region — same API, base URL routed automatically
ByteDance
$0.0000
$0.0000
0K
92
50
+60%
Seed3D 1.0 (International)NEW3D · Intl
Seed3D 1.0 international — coming when intl key is configured
ByteDance
$0.0000
$0.0000
0K
85
55
+40%
Kimi K2.7 Code HighspeedNEWFast
Same K2.7 Code quality with ~180 tok/s output
Moonshot
$0.950
$4.00
256K
90
92
+18%
Kimi K2.5
Proven stable release — strong at coding and Chinese tasks
Moonshot
$0.300
$1.50
131K
83
73
-5.2%
Grok Build 0.1NEWCode
Agentic coding model (replaces grok-code-fast-1)
xAI
$1.00
$2.00
256K
88
85
+12%
MiniMax M2.7NEWCatalog
MiniMax latest — catalog listed; API routes when MINIMAX_API_KEY is configured
MiniMax
$0.300
$1.20
205K
88
79
+12%
MiniMax M2.7 HighspeedNEWFast
M2.7 highspeed variant — ~100 tok/s
MiniMax
$0.300
$1.20
205K
88
90
+10%
GLM-5
Legacy alias — routes to GLM-5.1
Zhipu AI
$1.00
$4.00
200K
91
74
+8.5%
Prices shown per million tokensQuality/Speed scores from AMA composite benchmark, updated weeklyPrompt caching: save up to 90% on repeated inputsBatch API: 50% discount on async workloads