Pick the right model. Pay the right price. Switch anytime.
14 models routed through a single API. Pricing pulled straight from providers plus a transparent markup — no "unlimited" claims with hidden throttles.
One message format. Fourteen models.
Provider credentials stay server-side.
Unsupported models fail with typed API errors.
Streaming chunks preserve OpenAI-compatible shape.
Usage events capture tokens, provider, model, latency, and cost.
Every model, ready for production.
Prices shown include a 30% markup on top-tier models or 50% on cheap models vs publisher rates. Synced from models.dev on every build.
GPT-4o
gpt-4o-2024-08-06
Đa năng, multimodal, độ trễ thấp cho production app.
Input
$3.25
325 credits
Output
$13.00
1300 credits
Claude Opus 4.7
claude-opus-4-7
Smartest model của Anthropic. Reasoning sâu, code quality cao.
Input
$6.50
650 credits
Output
$32.50
3250 credits
Claude Sonnet 4.6
claude-sonnet-4-6
Sweet spot price/intelligence cho production agent.
Input
$3.90
390 credits
Output
$19.50
1950 credits
DeepSeek V4 Pro
deepseek-v4-pro
Frontier reasoning của DeepSeek. Code & math top-tier, giá thấp hơn Opus 3 lần.
Input
$0.570
57 credits
Output
$1.13
113 credits
Gemini 2.5 Pro
gemini-2.5-pro
1M context, multimodal đầy đủ. Đọc video/audio/PDF native.
Input
$1.63
163 credits
Output
$13.00
1300 credits
Mistral Large 2
mistral-large-2411
EU-hosted, GDPR-native. Tool calling chắc tay.
Input
$2.60
260 credits
Output
$7.80
780 credits
Kimi K2.6
kimi-k2.6
Moonshot Kimi K2.6 — agentic + reasoning, 262K context, mạnh ở task dài.
Input
$1.23
123 credits
Output
$5.20
520 credits
GLM-4.6
glm-4.6
Z.AI GLM-4.6 — bilingual zh/en, code/agent tốt, giá sweet spot.
Input
$0.590
59 credits
Output
$1.95
195 credits
Llama 3.1 70B
llama-3.1-70b-instruct
Open weights. Cost-effective qua Together/Groq/Fireworks.
Input
$0.940
94 credits
Output
$0.940
94 credits
Qwen 2.5 72B
qwen2.5-72b-instruct
Mạnh nhất dòng open Asia. Code & math tốt.
Input
$0.450
45 credits
Output
$0.520
52 credits
GPT-4o mini
gpt-4o-mini-2024-07-18
Rẻ nhất, đủ thông minh cho classify, extract, mass workloads.
Input
$0.220
22 credits
Output
$0.900
90 credits
o3 mini
o3-mini-2025-01-31
Reasoning step-by-step giá hợp lý cho code & STEM.
Input
$1.43
143 credits
Output
$5.72
572 credits
Claude Haiku 4.5
claude-haiku-4-5
Nhanh nhất, rẻ nhất trong dòng Claude. Latency thấp.
Input
$1.50
150 credits
Output
$7.50
750 credits
Gemini 2.5 Flash
gemini-2.5-flash
Multimodal nhanh, rẻ, thay thế cho gemini-2.0-flash đã EOL.
Input
$0.450
45 credits
Output
$3.75
375 credits
Mistral Small 3
mistral-small-latest
Open-weight, chạy được local. Cheap multilingual.
Input
$0.220
22 credits
Output
$0.900
90 credits
DeepSeek V4 Flash
deepseek-v4-flash
Reasoning fast tier rẻ kinh khủng. 1M context, $0.14 input/1M.
Input
$0.210
21 credits
Output
$0.420
42 credits
Pricing snapshot is synced from models.dev (MIT, maintained by SST) — run `pnpm sync:models` monthly.
General reasoning, multimodal chat, and tool calling for production apps.
Strong code quality, deep reasoning, 1M context, prompt caching to keep costs down.
Native multimodal (video/audio/PDF), 1M context, low latency.
DeepSeek + Moonshot + Z.AI — strong reasoning, cheap, long context 200K-1M.
EU-hosted, open-weight, self-hostable via the OpenAI-compatible adapter.