01 / Unified
One API. Every model.
Swap between GPT, Claude, Gemini, DeepSeek, Llama and 50+ models without rewriting a single line. OpenAI-compatible by default.
OneAPIforeverymodelyourstackwillneed.
›
The unified inference layer for GPT, Claude, Gemini, DeepSeek, Llama and 50+ models. Transparent pricing, predictable latency, one key.
API live · OpenAI & Anthropic compatible
OpenAI
GPT-5.4
Input · per 1M tokens
Official$1.25
TokenLayer$0.25-80%
Output · per 1M tokens
Official$10.00
TokenLayer$2.00-80%
Effective savings: up to 80% across all models
8 models · Auto-rotating · Use arrows or dots to navigate
01
Max savings vs. going direct
0%
02
Models in catalog
0+
Capabilities
Theinfrastructurelayerbetweenyouandeverymodel.
Everything you need to ship reliable AI products — without managing six provider contracts.
How it works
Three steps from signup to scale.
est. setup time
≈ 4 minutes
01
Create your account.
Sign up in 60 seconds, no credit card. Claim $20 of free credits to try every model.
› signup —› tokenlayer.net/signup
02
Grab your API key.
Generate a key, set a monthly cap, and we'll handle quota across every provider.
› TOKENLAYER_KEY=tl_live_a1b2c3d4...
03
Start building.
Drop-in compatible with the OpenAI SDK. Change baseURL — keep your existing code.
› client.chat.completions.create({ model: "gpt-5.4" })
The model catalog
Every frontier model. Always up to date.
New models land on TokenLayer shortly after release — same SDK call, same billing, same observability.
The model catalog
Every frontier model. Always up to date.
New models land on TokenLayer shortly after release — same SDK call, same billing, same observability.
catalog updated continuously
OpenAI
GPT
Flagship reasoning and multimodal models.
5.45.1o3
model.gpt
Anthropic
Claude
Long-context and coding-grade quality.
Opus 4.8Sonnet 4.6Haiku 4.5
model.claude
Google
Gemini
Native multimodal with massive context windows.
3 Pro2.5 Pro2.5 Flash
model.gemini
DeepSeek
DeepSeek
Frontier-grade reasoning at a fraction of the cost.
V3.2R1V3
model.deepseek
Meta
Llama
Open-weight models optimized for speed.
4 Maverick4 Scout3.3 70B
model.llama
+ 50 more
And more
Every notable provider, one endpoint.
MistralQwenxAICohere
model.andmore
+ 50 more
New models, soon after they ship.
Frontier models are added shortly after release, with streaming, tools and structured outputs supported throughout.
SCROLL
7 MODELS
Calculate your savings
See how much you'll save.
Adjust the sliders to estimate your monthly costs based on real provider pricing.
Tip: Drag the sliders or click any value to type your own
Select Model
1001.0M
10032,000
Total tokens / mo
20.0M
OpenAI
$1.25/$10 / 1M
Your monthly cost
Direct from provider
$95.00
With TokenLayer-80%
$19.00
You save / month
$76.00
You save / year
$912
Estimates based on 60/40 input/output split. Prices from pricepertoken.com
Developer experience
Fiveminutestoyourfirsttoken.
OpenAI and Anthropic compatible API with native streaming, structured outputs and tools. Keep your SDK — change one line.
- Works with the official OpenAI SDK
- Anthropic Messages API supported too
- Streaming + structured outputs out of the box
- Per-request usage & cost in your dashboard
app.ts
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
// Official OpenAI SDK — one line changed. import OpenAI from "openai"; const client = new OpenAI({ apiKey: process.env.TOKENLAYER_KEY, baseURL: "https://api.tokenlayer.net/v1", }); const res = await client.chat.completions.create({ model: "claude-sonnet-4-6", messages: [{ role: "user", content: "Summarize Q3." }], stream: true, }); for await (const chunk of res) { process.stdout.write(chunk.choices[0]?.delta?.content ?? ""); }
● 200 OK→ claude-sonnet-4-6$0.0021
642ms · 1,284 tokPricing
Transparent.Nosurprises.
No subscriptions, no seat fees. Top up, get multiplied credits, and pay per token.
Pay-as-you-go · no lock-in
Models are metered at provider list prices — your topped-up credits are simply worth more. Estimate your savings →
What you can do
Buildmore.Spendless.Sleepbetter.
Concrete outcomes you can expect after pointing your stack at TokenLayer.
Up to 80% below provider list prices
Questions
Common questions, honest answers.
Still curious? Ping us on Telegram — it's the fastest way to reach the team.
Yes. Point your base URL at api.tokenlayer.net/v1 and use any model in the catalog. Our endpoints are compatible with the OpenAI Chat Completions API and the Anthropic Messages API — including streaming, tool calling, and structured outputs.
Ready when you are
Buildwitheverymodelbeforelunch.
Sign up and claim $20 of free credits. No credit card required.
No credit card required$20 free creditsOpenAI & Anthropic compatiblePay-as-you-go