G
YouTube TikTok Games
NewClaude Fable 5.2 is live

One API for
every leading model.

Better prices, better uptime, no subscription. Prices and live status are public, and usage settles in USD.

Create a free accountRead the docs
Tokens, last 7 days
25.2B
Platform-wide usage, updated daily-39% on the previous 7 days
Request success, 24h
97.50%
All public models, weighted by requestsRequests fail over when a route has trouble
Public models
49
Live prices and status
Model providers
12
One account for all of them
The life of a request

From one base_url line
to a hard quota stop

  1. 01

    Change one line of base_url

    Request bodies, streaming and error shapes match the upstream. All three protocols are served at once, so existing clients only change one address.

  2. 02

    Fail over to a healthy route

    Upstreams are health scored and switched automatically: when one route fails, the request finishes on another.

  3. 03

    Stop when the quota runs out

    Every key and every team member can have its own quota and expiry. Usage stops at the limit and never overdraws quietly.

from openai import OpenAI client = OpenAI(- base_url="https://api.openai.com/v1",+ base_url="https://nextco.work/v1",  api_key="sk-cw-…",)client.chat.completions.create(  model="deepseek-v4.1-flash",  messages=[{"role": "user", "content": "hello"}],)
  • OpenAIChat CompletionsPOST /v1/chat/completions
  • OpenAIResponsesPOST /v1/responses
  • AnthropicMessagesPOST /v1/messages
Health scoring and automatic failoverGatewayRoute AHealthyRoute BHealthyFailing · switchedRoute CHealthy
97.50%24-hour request success rate
A quota and expiry for every key
Production$42.10 / $100
Staging$8.80 / $10
CI botStopped$20 / $20
Billing

Transparent prices,
a bill you can recompute

Input, output and cache hits are priced separately, and every record keeps the rate snapshot from call time. Price changes never rewrite past bills.

  1. Input5,404 ÷ 1,000,000 × $0.15 × 0.25 = USD 0.000203
  2. Cache hits12,800 ÷ 1,000,000 × $0.003 × 0.25 = USD 0.00001
  3. Output1,936 ÷ 1,000,000 × $0.60 × 0.25 = USD 0.00029
  4. Total3 lines added up = USD 0.000503Every record keeps the rates in force at call time

DeepSeek V4.1 FlashSample call

Input5,404 tokensUSD 0.000203

Cache hits12,800 tokensUSD 0.00001

Output1,936 tokensUSD 0.00029

TotalUSD 0.000503

Featured models

The most used models this week

Prices and status are read live. Keep scrolling and the shelf moves sideways.

All 49 models

DeepSeek V4.1 Flash

deepseek-v4.1-flash

#1
75% offGift balance acceptedTime-of-day pricing
Input
$0.0375$0.15
Output
$0.15$0.60

Input / output · per million tokens

Tokens / wk
9.1B
First token
3.7s
Weekly growth
-23%

DeepSeek Flash

deepseek-v4.1-flash-1017

#2
20% offGift balance acceptedTime-of-day pricing
Input
$0.12$0.15
Output
$0.48$0.60

Input / output · per million tokens

Tokens / wk
7.7B
First token
7.0s
Weekly growth
-60%

Claude Opus 5.5

claude-opus-5-5

#3
84% offCash only
Input
$0.64$4.00
Output
$3.20$20.00

Input / output · per million tokens

Tokens / wk
2.3B
First token
8.6s
Weekly growth
-65%

Claude Sonnet 5.5

claude-sonnet-5-5

#4
74% offGift balance accepted
Input
$0.52$2.00
Output
$2.60$10.00

Input / output · per million tokens

Tokens / wk
2.2B
First token
3.6s
Weekly growth
+41%

Claude Fable 5.2

claude-fable-5-2

#5
New84% offGift balance accepted
Input
$1.28$8.00
Output
$6.40$40.00

Input / output · per million tokens

Tokens / wk
1.8B
First token
9.6s
Weekly growth
New

GPT-6.1 Sol

gpt-6.1-sol

#6
74% offGift balance accepted
Input
$0.52$2.00
Output
$2.60$10.00

Input / output · per million tokens

Tokens / wk
741.4M
First token
11.1s
Weekly growth
+15%

GPT-6 Luna

gpt-6-luna

#7
74% offGift balance accepted
Input
$0.026$0.10
Output
$0.13$0.50

Input / output · per million tokens

Tokens / wk
621.4M
First token
9.5s
Weekly growth
-22%

Claude Haiku 5.5

claude-haiku-5-5

#8
74% offGift balance accepted
Input
$0.026$0.10
Output
$0.13$0.50

Input / output · per million tokens

Tokens / wk
418.5M
First token
2.0s
Weekly growth
+154%

Rankings

156Btokens · last 28 days

Ranked by real token usage on the platform, so you can see what people actually run. The chart replays the last 8 weeks; drag the timeline to look back.

Full rankings
  • DeepSeek
  • Anthropic
  • OpenAI
  • xAI
  • Other vendors

Model usage ranking, rolling 7-day totals
ItemSep 28Sep 30Oct 2Oct 4Oct 6Oct 8Oct 10Oct 12Oct 14Oct 16Oct 18Oct 20Oct 22Oct 24Oct 26Oct 28Oct 30Nov 1Nov 3Nov 5Nov 7Nov 9Nov 11Nov 13Nov 15
GPT-6 Luna2.5B2.2B1.7B1.1B532.8M142M242.7M275.8M419.8M1.4B1.5B1.6B1.2B428.5M444.6M499.5M626.3M782.3M947.3M799M796.6M472.8M407.4M486.1M687.5M
Claude Opus 5.51.7B1.5B1.1B930M277.3M1.4B1.9B2.1B1.9B941.4M775.8M3.8B4B6.7B6.8B3.3B1.3B758.9M4.5B6.6B6.6B6.4B2B2B2.5B
GPT-6 Sol1.2B1B374M326.5M125.6M63.6M20M18M33M47.8M51.5M50.6M38.7M24.5M4.8M17.1M24M19.2M19.2M181818000
DeepSeek-V4-Flash708.5M145.6M152.6M62.8M84.6M118.7M56.1M56M4.7M4.7M4.7M47.3M59.3M240.4M253.9M196.5M142.7M13.3M0569.8K2M2M2M00
Claude Opus 5298.1M201.3M64644826.8K26.7K26.7K0355.2K887.6K887.6K1.1M180.6K180.7K180.2K2002000000000
Claude Fable 5.1451.5M5.2M5.2M5.2M131.7M134.9M136.4M136.5M13M8.9M8.9M858.6K8.4M7.8M9.5M8.4M1.9M1.9M2.2M6.6M10.9M11.1M5M11.3M13.6M
Grok 4.6247.7M78.4M73.9M71M17.2M4.6M23.9M29.3M29.3M27.4M8.8M5.3M11.6M10.2M8.9M2.4M1.2M540.4K62.2K60.5K21.5K21.5K000
GLM-5.3-Flash268.9M61.8M00000015.6M15.7M15.8M15.8M1.9M1.9M49.5M47.6M47.6M47.6M05.5M6M7M2.2M1.5M1.5M
MiMo-V2.6-Flash443M343M240.7M149.8M363.3K167.6K189.1K21.5K21.5K001.3M1.3M1.3M20.3K07M7M7M2424242400
GPT-6 Astra68.1M189M189.1M172.9M18.6M19.4M324.2M332.8M322.7M281.6M15.7M56.8M64.2M64.2M7.5M155.9K45.1K45.1K20M20M38.4M38.4M18.8M7.4M6.4M
MiMo-V2.6-Pro195.3M190M111.9M30M2.6M32.4M58.6M204.2M213.2M225M149.6M67.5M34.3M5.6K5.6K2.2K0002424242400
GPT-6.1 Sol074.3M524.5M649M910.2M815.2M781.4M719M665.4M713.6M1.3B1.4B1.5B997M625M227.9M274.2M269.1M257.6M324.3M646M797M1.1B871.3M612.8M
GPT-5.6 Sol85.3M125.7M121.5M73.4M03.9M21.3M21.6M22.2M30.3M53.6M48.9M33.4M38.4M36M41.9M5.9M5.9M5.9M1.5M3.6M3.6M129.4M126.9M126.9M
Claude Sonnet 5.5084.6M143.9M143.9M95.9M227.1M316.9M931.3M1B1B499.5M808.3M2.1B2.4B2.7B710.1M491.3M227.9M555.5M1.2B1.6B1.8B1.1B1.1B1.6B
Grok 4.731.7M1.1M605.5K568.1M1.4B4.2B4.2B3.6B1.8B185.5M42.8M50.7M50.8M42.8M42.1M409.3K5.3M5.5M13.5M18.9M13.6M47.3M67M67.4M67.4M
Claude Haiku 5.500000215.4M1.4B1.5B1.8B1B555.6M347.3M725M730.3M721.7M68.1M54.3M46.3M173.4M183.3M165.1M18.2M4.1M418.8M425.7M
DeepSeek-V4-Flash-Vision-Exp00000002.8M845.5M845.5M842.7M132.1M154.3K154.3K154.3K153.7K0004747474700
deepseek-v4.1-flash-ultrafast00000000166326.2M345.3M701.8M663.5M374.1M90.4M0000000000
DeepSeek Flash0000000000294.7M7.2B18.2B25.5B27.6B21B19B19.6B22.1B23.7B19.3B15.3B12.6B8.1B8B
Claude Sonnet 565.4M41.1M33.2M34.6M34.6M6.8M1.1K0000699.1M1.2B1.7B1.8B660.9M343.5M89.2M382.6M432.9M423.5M188.4M000
minimax-m3.1-flash00000000000557.6M1.4B1.5B1.7B675.4M362M303.7M0000000
DeepSeek V4.1 Flash0000000000001.5B8.6B17.7B21.1B18.9B13.2B9.6B10.1B11.7B11.4B10.5B9.7B9.7B
DeepSeek-V4-Pro128.3M124.9M120M80.5M19647.1K46.9K46.9K47.8K8758752332.3K2.3K2.3K000910.5K70M100.5M119.6M181.8M141.4M141.4M
Claude Fable 5.200000000000000000000001.1B1.6B1.8B
NextCoWork desktop

One window, four layers working for you

An agentic coding desktop app. Open a local folder as your workspace andthe agent reads, searches and runs commands inside it.

  • L4Credentials never reach the renderer

    The renderer is a pure IPC client with no network or credential access; model requests go through the system network stack, so corporate proxies and certificates keep working.

  • L3Many upstreams, one shape

    Upstream routing normalises protocols, scores health and fails over automatically, so one dead provider does not end the conversation.

  • L2An agent loop that uses tools

    think → tool → observe, kept moving: file edits, code search, command execution, browser and web search, tasks and to-dos.

  • L1Your local folder is the workspace

    Open a directory and the agent reads, searches and executes inside it — path fencing keeps it from wandering out.

Download NextCoWorkmacOSWindowsLinuxv2.8.7 · released Nov 3, 2026 · Stable channel
Skills and plugins

Teach your agent more

Find workflows, writing helpers and automations shared by the community as Skills; plugins add editors, agent tools and UI to the desktop client.

Terminal — nextcowork
  1. Command: npx @aidotnet/nextcowork install skill programming-tutor. Downloading programming-tutor, complete. ✓ SHA-256 matches the server digest ✓ Installed to ~/.next-cowork/skills/programming-tutor
269 Skills · 13 pluginsBrowse SkillsBrowse plugins

FAQ

Frequently asked questions

A few things you might want to know before you start.

  • Creating an account is free, and model calls settle in USD by actual usage. Input, output and cache hits are priced separately per model, and some models add tiered or time-of-day rates. The model page always shows the price currently in effect. A new account does not include free call credit.

Read the docs

Change one base_url.
Start calling today.

Free to sign up, pay for what you use, stop any time.

Create a free accountRead the docs