building consumer apps @polldotfun, @miniroutersh, airdropped.link
ldn/nyc
- Tweets 14,584
- Following 4,953
- Followers 18,938
- Likes 156,470
Finally, introducing Offers to everyone.
Offers is like a P2P sportsbook.
1. Create a fixed odds offer on any football game
2. Anyone can take the other side, partially or fully
3. Resolves automatically, funds settle instantly.
Introducing Offers on Poll.
Pick any major football game, set your own odds and put the money up.
Anyone can take you on, all of it or just a slice.
Paid out automatically at full time.
2
4
21
4,175
Zhe retweeted
GPT-6.1 Sol is now live on minirouter.sh.
>80% cheaper than GPT-6 Astra
>nearly as capable as Astra
>rank #4 in intelligence index
>replaces GPT-6 Sol just after 7 days
GPT-6.1 Sol: near-Astra intelligence for a fifth of the price.
It’s the most cost-efficient model for its performance available today.
9
13
57
3,514
Zhe retweeted
SUPERNOVA WINNER: @miniroutersh by @iamzhe
minirouter lets you use different AI models through one API and pay with crypto.
switch models whenever you need to, without opening another account or managing another balance.
340 accounts created, 39 funded, and topups grew 50% last week.
builders will win. meteora.
10
25
70
4,849
Zhe retweeted
More Minirouter updates today 🚢
1. minirouter:web_search is live in alpha. One cent a search, powered by @perplexity_ai.
2. Our agent swarm library is out. First use case: cheap, parallel code reviews.
3. BTC and ETH deposits now show live status, with a tracking link.
4. Usage and cost export for personal and team accounts.
19
21
85
8,471
claude making an absolute comeback with models
still finding efforts strange though
>best effort for opus 5.5 = medium
>best effort for sonnet 5.5 = xhigh
Sonnet 5.5 is now available on minirouter.sh.
1. 30% faster and cheaper than Sonnet 5.
2. As smart as GPT-6 Astra (max) 🤯
3. Ranks #4 in AA's Intelligence Index.
2
25
2,580
Zhe retweeted
new ATH in terms of tokens served
@miniroutersh is killing it
ok more than doubled the amount of tokens served by @miniroutersh merely 5 days later
@miniroutersh also just had its highest usage day
bullish
minirouter.sh/analytics
1
11
34
4,852
Zhe retweeted
Introducing Self-Serve for Agents.
1. One POST returns an API key and a deposit address.
2. Pay in USDC, SOL or any SPL token.
3. Access 240+ models, no human needed.
minirouter.sh/changelog/agen…
7
10
64
4,325
Zhe retweeted
ok more than doubled the amount of tokens served by @miniroutersh merely 5 days later
@miniroutersh also just had its highest usage day
bullish
minirouter.sh/analytics
6
8
23
3,794
Zhe retweeted
Introducing Teams on minirouter.sh.
- One shared balance for up to 10 people.
- Every member gets their own API keys.
- Owners set limits, guardrails and routing.
- Members can share presets.
Team bigger than 10? DM us and we'll get you sorted.
13
21
84
4,153
>10x smaller and more accurate than Jev btw
Introducing GLiNER2.5-Decide, our new 340M parameter open weight, encoder-based decision model.
GLiNER2.5-Decide is built for fast, deterministic classification. The model evaluates a set of user-defined typed questions and rules, and jointly decodes their answers, returning structured decisions with probability distributions and confidence scores.
We evaluated the model’s performance on Fast Decisions, an unseen, internally generated classification suite based on 17 datasets testing real-world use cases across routing, triage, classification, sentiment, and content understanding.
Measured against similar decision models, GLiNER2.5-Decide leads in 9 of the 17 datasets, achieving the highest average score:
- GLiNER2.5-Decide: 60.1%
- SemIf: 56.4%
- JevK5: 57.5%
- Laya: 46.6%
This performance makes the model a strong fit for use cases like tool calling, model routing, browser and computer use, and LLM-as-a-judge.
GLiNER2.5-Decide’s lightweight encoder architecture makes it easy to fine-tune the model for specific tasks, while being efficient enough to run locally on consumer-grade CPUs or in air-gapped environments, giving users greater control over where their data goes and where the model runs.
To make building and experimenting with GLiNER2.5-Decide as easy as possible, we're also offering hosted inference. You can now use our API to run inference and fine-tune GLiNER models on specialized tasks right inside your own coding agent: agent.fastino.ai
As with previous models, we’re also releasing the model weights on @huggingface under the Apache 2.0 license: huggingface.co/fastino/GLiNE…
5
31
3,274
Zhe retweeted
meet the first SUPERNOVA INFRA teams:
@embercurve is building a next-gen, community aligned launchpad.
@miniroutersh puts hundreds of AI models behind one API, payable with crypto.
@PerpsPadfun connects token launches to live perpetual markets.
@getstonkoptions distributes tokenized stocks to employees.
@VaultBags turns trading fees into tokenized RWA rewards.
the SUPERNOVA INFRA initiative will continue to evolve.
stay tuned for more information on future awards.
54
60
186
34,132
also shipped a few fixes alongside these 🚢
> jev input no longer capped at 8KB
> if a response format can't be enforced, we now tell you
> documented the usage param quirks
> missing models are back in the catalog
Two new models are now live on Minirouter.
1. qwen3.8-max-prime from @Alibaba_Qwen.
2. ember-1 from @FireworksAI_HQ.
minirouter.sh/changelog/qwen…
3
6
31
2,728
Zhe retweeted
Pay with $MINI, get 5% extra credits.
Spend them across 240+ models.
Live now on minirouter.sh.
23
16
91
14,452
1st half of minirouter.sh's BYOK router is now live.
You bring multiple keys, we switch between them when limits are exhausted.
2nd half lets you bring your subscriptions, live soon.
Bring Your Own Key (BYOK) support is now live.
1. Add keys from different accounts.
2. We automatically switch between keys when limits are exhausted.
Supporting OpenAI, Anthropic and Deepseek today.
minirouter.sh/changelog/brin…
6
6
45
3,615
HUGE: @BlackRock just released an AI x Crypto thesis on INFERENCE CAPITAL MARKETS
These are the tokens on Solana and Robinhood you can research & buy:
"Compute is emerging as a new and potentially large market for digital assets."
Inference Capital Markets
@orbiodotso $72M
@miniroutersh $2.7M
@UsePodAI SQUIRE $1.6M
...but what even is Inference Capital Markets and why does it matter?
today, an inference provider has to buy GPUs, power, data center space, and much more, all of this upfront
with alternatives like Orbio or Squire, anyone can pre-sell or finance future capacity
the other side of this market, is that it allows for speculation. if you think inference will get more expensive, you can buy capacity today and sell in the future for higher if you are right, or just use it yourself
crypto is the perfect infrastructure for inference capital markets because it provides a programmable, transparent layer to coordinate these operations without relying on a single intermediary
tl;dr
- you can sell your unused ChatGPT/Claude/Gork credits
- you can buy credits much cheaper than using LLMs directly
- you can speculate on inference prices if you have an edge
Our latest research paper explores the growing connection between AI and digital assets and explains why broad AI adoption may drive new demand, utility and applications across the digital asset economy. blackrock.com/us/individual/…
25
16
129
16,870
Zhe retweeted
GPT-6 Sol and GPT-6 Luna are live on minirouter.sh.
Roughly 50% cheaper than their 5.6 predecessors.
Please welcome GPT-6 Sol and GPT-6 Luna to the GPT-6 universe.
GPT-6 Sol and Luna build on the advances behind GPT-6 Astra, bringing much of its strengths into faster and more affordable models to support work at scale.
We’ve also made caching and inference more efficient, and we’re passing the savings directly to you: 50% lower API prices for Sol and Luna compared with GPT‑5.6 promotional pricing.
3
7
37
3,841
Opus 5.5 medium outperforms high, xhigh AND max
strange times
Opus 5.5 is now available on minirouter.sh.
>30% faster than Opus 5
>40% cheaper than Fable 5.1
>on par with GPT-6 Astra and Fable 5.1
>ranks #1 in AA's Intelligence Index
4
1
43
6,258