UplinkSwitch

One API key for 83 AI models: Claude, GPT, Gemini, Grok, uncensored builds, Veo 3.1 and Nano Banana Pro. Up to 70% under list price. Prepaid, pay per token.

[UplinkSwitch @uplinkswitch]

Flagship model APIs, per million tokens, ours vs the developer's list price:

Claude Opus 5.5: $2.10 / $6 (list $4 / $20)
Claude Sonnet 5.5: $1.20 / $3 (list $2 / $10)
GPT-6.1 Sol: $0.60 / $3 (list $2 / $10)
Gemini 3.1 Pro: $1.20 / $7.20 (list $2 / $12)
Grok 4.7: $0.60 / $1.80 (list $2 / $6)
Kimi K3: $0.90 / $4.50 (list $3 / $15)

One key and one base URL for the OpenAI, Anthropic and Gemini SDKs, Claude Code and Cursor. Prepaid, pay per token, no subscription.

Every model is tested daily and the token counts are public: https://t.co/wQYdIEU6Ym

https://t.co/SDnyPdQldK

[UplinkSwitch @uplinkswitch]

No subscription. Top up $10 to $500 by card, Apple Pay, Google Pay or crypto, and the balance only moves when you send requests.

Prices lock for 30 days after each top-up. Failed image and video jobs are refunded automatically.

https://t.co/oGgWP9mCXv

[UplinkSwitch @uplinkswitch]

How to check whether your API provider bills tokens you never sent:

Send a tiny prompt with max_tokens set to 1 and read usage.prompt_tokens. A clean route reports a handful of tokens.

On 2026-10-12 we measured routes that reported 4,469 input tokens for a 90-token prompt, and two that reported 14,111.

The curl and the full numbers: https://t.co/p0J7zzKaet

[UplinkSwitch @uplinkswitch]

@cultured @sequoia @Shopify @jakepaul To the moon!!

[UplinkSwitch @uplinkswitch]

Kimi K3 at $0.90 / $4.50 per million tokens. Moonshot's list price is $3 / $15.

GLM 5.3, Qwen3.7 Max and Plus, MiniMax M2.7 and MiMo are on the same key, each under its developer's list price.

https://t.co/fTIGKDFdoV

[UplinkSwitch @uplinkswitch]

@spacedevin https://t.co/oGgWP9m57X

[Steven Schwartz 🦅 @cultured]

Whop has raised $200M in Series D funding at a $5.1B valuation, backed by @sequoia and @shopify.

It really feels like we are barely touching the tip of the iceberg in a market larger than anyone can imagine.

We first launched Whop as a better way for us to sell sneaker bots, one of the most niche types of desktop software out there.

Total TAM? Probably $20,000 per day.

Who cares, tiny market.

But, what people discounted is the idea that there were actually 70,000 other markets exactly like that one.

Beyond the investment itself, I am especially excited to welcome legendary silicon valley investor @dougleone to our board of directors. Doug and the sequoia team are nothing but a brilliant, resourceful, and principled group of people.

My co-founder Cameron and I started building on Shopify when we were 13 years old. Shopify is one of the most impactful all-in-one cloud ecosystems ever created; @tobi has pioneered modern entrepreneurship the way we know it and we are thrilled to have him in our corner.

We have a tenet at Whop to be the dumbest in the room and I am very grateful that this has recently become a lot easier.

We are focused on markets, scale economies, and progressive media. If that is up your alley, please message me directly.

The $200M raised will support our expansion as we build economic infrastructure for the next century, continuing on our mission of delivering everyone an income.

[Iman Gadzhi @GadzhiIman]

Generational company. Generational team.

I invest to be on the right side of history. Whop has been a home run, but what we're building matters more: a brighter future for builders and dreamers.

Proud of the team. We're just getting started.

[UplinkSwitch @uplinkswitch]

What $1 of API credit buys at 2,000 tokens in and 500 out per turn:

Claude Haiku 5.5: about 5,128 turns
Claude Sonnet 5.5: about 256 turns

New accounts get $1 of credit to try it: https://t.co/BJwHnqbNnA

[UplinkSwitch @uplinkswitch]

Grok 4.7 at $0.60 in / $1.80 out per million tokens. xAI's list price is $2 / $6.

Grok 4.6, 4.5, 4.3 and the 4.20 multi-agent build are on the same key.

https://t.co/tBNISaUigx

[UplinkSwitch @uplinkswitch]

Moving an app to UplinkSwitch is usually two env vars and no code change:

OPENAI_BASE_URL=https://t.co/mzS9feJIhy
ANTHROPIC_BASE_URL=https://t.co/mzS9feJIhy

The Google Gen AI SDK takes the same URL in http_options. The Vercel AI SDK works through @ai-sdk/openai-compatible.

https://t.co/ueDrCNwPir

[UplinkSwitch @uplinkswitch]

Nano Banana Pro at $0.0804 an image at 1K or 2K. Google's price is $0.134.

4K: $0.144 ($0.24)
Edit at 1K or 2K: $0.0804
100 images at 2K: $8.04 here, $13.40 at Google's price.

It runs through the standard OpenAI images call. No Google account or billing setup.

https://t.co/FhLBSpT32d

[UplinkSwitch @uplinkswitch]

DeepSeek V4 Pro at $0.792 / $1.92 per million tokens (list $1.32 / $3.96). V4.1 Flash at $0.06 / $0.18 (list $0.30 / $1.20).

There's also an uncensored V4 Pro build at $0.75 / $0.75, on the same key.

https://t.co/Z4LCO3IBhl

[UplinkSwitch @uplinkswitch]

Cline, Roo Code, Kilo Code and OpenCode all take an OpenAI-compatible provider.

Base URL https://t.co/d6LePgDTQs, your key, a model id. Claude, GPT, Gemini, Kimi and the uncensored builds then sit in the same model list.

https://t.co/Gs7jIiOGFX

[mirku @mirku21]

THIS IS A ACTUAL F*CKING GOLDMINE

Anthropic just shared one of the best tricks for Claude Code

The idea: run Sonnet, Haiku 5.5 and Opus 5.5 at once, each doing only what it's best at

Better results and way fewer tokens: every task lands on the right model

You just start Claude Code like this:

CLAUDE_CODE_SUBAGENT_MODEL=haiku claude --model sonnet

here's how it splits:

→ Sonnet 5.5 runs the main session, writes the code, and applies fixes
→ Haiku 5.5 swarms the repo in parallel: search, grep, docs, and test audits at 340 tok/s
→ Opus 5.5 stays on call as architect for the big decisions and signs off on diffs
→ you stop burning Opus context on routine tasks and stop flooding main session with raw logs

Opus only steps in at three moments:

→ before a plan locks
→ when the same error hits twice
→ before calling it done (signing off on git diff)

Sonnet builds, Haiku swarms, Opus reviews

JEV engineering is the same move one layer down: the forks that need no thinker (which file, which tool, retry or stop) go to JEV in milliseconds, and the big models only see the ones that split

the full tree
> Sonnet 5.5 runs the main session and coordinates code implementation
> search maps the repo on Haiku 5.5 (read-only, 340 tok/s)
> grep finds call sites on Haiku 5.5 (0.05x cost, isolated context)
> audit scans diffs in a temporary worktree (isolation: worktree)
> Opus 5.5 on call as architect, reviews synthesized diff before sign-off
> JEV routes micro-forks in milliseconds

paste the tree and this prompt into Claude Code ↓

"Reconfigure my Claude Code setup around this tree:
1. Run the main session on Sonnet 5.5 with Opus 5.5 as architect
2. Create three custom Haiku 5.5 subagents in .claude/agents/: search (maps repo), grep (finds call sites), and audit (runs isolated diff checks in a worktree)
3. Set isolation: worktree on audit so modifications never touch my working branch
4. Add one rule to CLAUDE.md: delegate all repository searches and test runs to Haiku 5.5 subagents to keep main context under 10k tokens
5. Route fast deterministic forks (tool choice, path selection, retries) through JEV
Show me every change as a diff first. No edits until I say go"

↳ https://t.co/3SePgfMHtW

save it, it'll come in handy

[UplinkSwitch @uplinkswitch]

Veo 3.1 on one API key, per 8-second clip (Google's price in brackets):

Veo 3.1 720p: $1.80 ($3.20)
Veo 3.1 1080p: $1.92 ($3.20)
Veo 3.1 4K: $2.88 ($4.80)
Veo 3.1 Fast 720p: $0.48 ($0.80)
Veo 3.1 Lite 720p: $0.24 ($0.40)

Failed clips are refunded automatically. Same key as Claude, GPT and Gemini.

https://t.co/Ou5G33l69B

[UplinkSwitch @uplinkswitch]

Gemini 4 Argon is $0.60 in / $3.00 out per million tokens here. Google's list price is $2 / $10.

It works in the Google Gen AI SDK with a base_url in http_options, or in any OpenAI-compatible client. Model id: gemini-4-argon

https://t.co/ie2QAkQtou

[Kevin Kern @kevinkern]

the release of opus 5.5 has been the biggest unlock for me so far when it comes to running multi-agent orchestration nonstop across different projects.

even codex benefits from it, since opus manages my gpt sol workers without burning through my usage limits within 6 hours by running astra as the main agent.

[UplinkSwitch @uplinkswitch]

Six uncensored and abliterated builds on one OpenAI-compatible key, one page:

glm-5.3-flash-uncensored $0.084 / $0.084 per 1M tokens
glm-5.3-uncensored $0.18 / $0.18
roleplay-uncensored $0.30 / $0.60
glm-5.3-flash-abliterated $0.66 / $0.66
deepseek-v4-pro-uncensored $0.75 / $0.75
glm-5.3-abliterated $2.25 / $2.25

Each one's daily token-count check is on the page. https://t.co/0A9I6SDkec

[UplinkSwitch @uplinkswitch]

GPT API prices vs OpenAI's list, per million tokens:

GPT-6.1 Sol: $0.60 / $3 (list $2 / $10)
GPT-6 Astra: $3 / $15 (list $10 / $50)
GPT-5.5: $1.50 / $9 (list $5 / $30)

The OpenAI SDK needs one env var: OPENAI_BASE_URL=https://t.co/d6LePgDm0U

https://t.co/CzXp3eGubg

[UplinkSwitch @uplinkswitch]

Janitor AI on uncensored models, one key, no per-site account.

In Janitor's API settings, use a Custom (OpenAI) proxy: base URL https://t.co/mzS9feKg76, your key, then a model id:

roleplay-uncensored: $0.30 in / $0.60 out per 1M tokens
deepseek-v4-pro-uncensored: $0.75 / $0.75
glm-5.3-abliterated: $2.25 / $2.25

All three stream. Prepaid from $10, no subscription.

Setup: https://t.co/9ZIpdpgC7M