@concentrateai
Account based inUnited States
About this account
- Account based in
- United States
- Connected via
- Web
Account-level information from X, not a live location or the device used for a specific post.
The governance platform and reliable AI gateway for production. One API for every model, no fees, built for enterprises and high-growth startups.
New York, NY
Joined May 2025
- Tweets290
- Following0
- Followers1.8K
- Likes579
Pinned Tweet
We’re giving away $25,000 to build with Gemini!
We have $25K in Gemini credits to give away, and we want it to go to a startup building something incredible.
Tell us what you’re building, what you’d use Gemini for, and why $25K of model usage would make a difference.
Apply here by 9/6 at 11:59pm PST: forms.gle/etcH3WQLrM3D8M8M8
Concentrate AI retweeted
I know @OpenRouter does more than us overall but pretty sick to see this today nonetheless...
SAVE 35% on GLM 5.3 on Concentrate via @novita_labs!
Now LIVE!
concentrate.ai/models/glm-5.…
Customers have used 100B DeepSeek V4.1 Flash tokens from @novita_labs over the last 24 hours at an avg of 213 T/S. ⚡️ fast!
Try here: concentrate.ai/models/deepse…
Concentrate AI retweeted
We're doing something special in AUSTIN Oct 4–10!
I rented a MANSION 17-bed house just outside Austin and I've invited 17 founders & builders from across the country to move in for the week.
Coworking by day. Private dinners downtown, special guests, BBQs, poker, yoga, and pool/hot tub by night. Nothing else like it.
Sponsored by @obvious , @concentrateai, and @frame_payments + one more big name to be announced soon.
7 beds left. Filling fast!!
Apply: luma.com/0l4a3kvp
WEEKEND SHIPPING!
We now have 6X the GLM-5.3-Flash providers that we had a week ago. This is one of the most popular models we've ever seen.
@FireworksAI_HQ and @CrusoeAI are now LIVE!
Check it out: concentrate.ai/models/glm-5.…
DeepSeek V4.1 Flash is live on Concentrate.
DeepSeek’s fast, multimodal agent model pairs million-token context with a new architecture built for higher throughput and lower inference cost.
- 1M-token context
- 552B MoE, 8B active on input / 16B on output
- Native image + text understanding
- 90.6 Terminal-Bench 2.1, 74.2 DeepSWE v1.1
Route DeepSeek V4.1 Flash and 175+ other models now: concentrate.ai/models/deepse…
Concentrate AI retweeted
50% SOLD OUT!
I rented an INSANE 17-bed house in Austin for founders for 7 days (Oct 4th-10th). Technical founders preferred.
Coworking. Private cocktail parties. 5-star dinners that sponsors are paying for.
Apply before the last spots disappear: luma.com/0l4a3kvp
SOLD OUT.
See you tonight.
Join the waitlist: luma.com/57tzm8qr
This was a LOT of fun
@CalBlockchain had a great time this weekend running Berkeley's Robotics Hackathon! 🦾
Some project highlights:
> Robots emulating actions from egocentric video
> Post-trained VLAs for automated food consumption
> Raspberry Pi based rover for monitoring plant health
Big thank you to all attendees and our partners World Context, @ElevenLabs, Praxis Robotics (YC S25), @Lovable, @concentrateai, and the Berkeley Haas Entrepreneurship Program.
Check out the projects below 👇
Muse Spark 1.3 is live on Concentrate.
Meta’s latest Muse model pushes further on coding and agentic tasks, with its biggest reported jump yet for long-running software and agent workflows.
- 61 on the Artificial Analysis Intelligence Index, tied with GPT-5.6 Sol and Grok 4.6
- Up from 57 for Muse Spark 1.2
- Same pricing as the previous Muse Spark release
Use Muse Spark 1.3 and 150+ other models now: concentrate.ai/models/muse-s…
BYOK is NOW FREE!
Have free credits but experiencing outages or 429s?
Fallback instantly from OpenAI to OpenAI hosted on Azure and keep your product up and reliable.
We always use your key first and then fallback to our own.
Gemini 3.8 Flash is live on Concentrate (the only gateway with no fees.
Google’s new Flash model improves coding, agents, and complex reasoning while staying ~the same speed and the same price as 3.7 Flash.
- 1M-token context
- $0.75/MTok input, $3.75/MTok output
Use Gemini 3.8 Flash and 175+ other models now: concentrate.ai/models/gemini…
Claude Fable 5.1 is live on Concentrate on 3 providers!
@AnthropicAI's model for demanding reasoning and long-horizon agentic work, with stronger coding, research, and complex knowledge workflows.
- 1M-token context, 128K max output
- Adaptive thinking, always on
- Stronger long-running agentic coding and multistep research vs. Fable 5
- $10/MTok input, $50/MTok output; $0.25/MTok cache reads
Use Claude Fable 5.1 and 175+ other models now: concentrate.ai/models/claude…
Concentrate can redact sensitive information before ever reaching the LLMs. Included for free. Or use our API for cheap!
concentrate.ai/features/data…
Concentrate AI retweeted
INSANE! OpenRouter collected $50M in fees last year.
A lot of that came from YC companies running their inference through them.
We’ve had more YC companies switch to @concentrateai in the last two weeks than in our lifetime before that.
If that’s you, DM me. I’ll personally onboard you and buy your team dinner.
Is your AI stack just OpenAI and Anthropic?
A lot of teams we talk to start like this.
You can use open-source models on ZDR providers, or try any model on @concentrateai with multiple providers behind each one.
Try for free at concentrate.ai
Had a lot of fun sponsoring @byhackathonco's event yesterday as the official inference provider.
If you want an AI Gateway with no fees check us out.
luma.com/k6owftb3
We're building the B2B OpenRouter.
INSTANTLY SAVE 5.5% on all AI spend vs OpenRouter.
Venture backed? Reach out for our intro offer.
concentrate.ai/
Concentrate AI retweeted
We're hosting a private poker night in SF
September 8th
Founders + engineers
@concentrateai x @runpentagon
Expect strong players
Apply:
luma.com/57tzm8qr
Concentrate is hosting a private founder poker night in SF with @runpentagon in 10 days.
Space is very limited.
Request a spot: luma.com/57tzm8qr?utm=cnaix