BitRouter

cost-optimize your agentic workloads | @kelrouter & @arcrouter | works with any harness, open sourced at

[OpenAI @OpenAI]

GPT-5.6 Sol, along with Terra and Luna, will launch publicly this Thursday.

We’re expanding preview access globally now.

[BitRouter @BitRouterAI]

autonomous sovereign agent need open models & open tools

all @agentkeyapp users get 25% off from all open-sourced models from BitRouter

[AgentKey @agentkeyapp]

We met @kelrouter, the founder of @BitRouterAI at @ETHGlobal NYC and were immediately impressed by his vision for open source AI infrastructure.

The future of AI won't be built by a single model or a single tool. It'll be powered by an open ecosystem where every component works together. That's why we're excited to announce our partnership with BitRouter.

Throughout July, all paid AgentKey users receive 25% off BitRouter 🎉

Here's to building faster, smarter, and more reliable AI agents together.

[kel @kelrouter]

If they can target Chinese users today, they can target anyone tomorrow — under the will of the US government.
There is no privacy with closed-source models.
Use open models — GLM-5.2, Kimi-K2.7-Code, MiniMax-M3 — on @BitRouterAI.
Defend your freedom.

[BitRouter @BitRouterAI]

@sagapad_icm @whoisetimfon Salute

[SagaPad @sagapad_icm]

7/ #ColosseumReview Race

🏆 Top 3 Overall Reviews

1st Place is @whoisetimfon

https://t.co/hU3jdFpB1P

This application won us over because @whoisetimfon didn't pick 1 track and call it a day but covered every single track, which means going through all of Colosseum's projects one by one, not just skimming. No surface-level shoutouts. Each project got a real, detailed review.

Colosseum projects got reviewed: @traded_gg, @BitRouterAI, @BreathProtocol, @xplaceapp, @refihub

[Movez @0xMovez]

https://t.co/OMBzNQETKM

[BitRouter @BitRouterAI]

agent spawn coming soon

[kel @kelrouter]

3 consecutive @ETHGlobal wins
build on consecutive track for @BitRouterAI with @unluckiXD → 1st prize from @SuiNetwork @WalrusProtocol

any agent can spawn subagents, pick models & pay autonomously — harness & model agnostic, built for long-run coding tasks.

releasing soon

[BitRouter @BitRouterAI]

you're paying frontier prices for calls that don't need frontier models.

kimi-k2.7-code is now live on BitRouter — with 25% off.

outperforms claude-opus-4-8 on tool use. costs 5x less.

open-source router. zero harness changes. optimizes cost & performance - with every run.

[Kimi.ai @Kimi_Moonshot]

🌘 Kimi-K2.7-Code, our latest coding model, is now released and open-sourced!

🔷 Improved coding & agent performance over K2.6: +21.8% on Kimi Code Bench v2, +11.0% on Program Bench, and +31.5% on MLS Bench Lite.
🔷 Reasoning efficiency: Less overthinking, with 30% lower reasoning-token usage compared to K2.6.
🔷 Long-horizon coding: Improved instruction following, higher end-to-end coding task success rates.

⚡️ 6x High-Speed Mode coming soon!
🔌 Available today via Kimi API and Kimi Code.

🔗 Kimi Code: https://t.co/uvoSJKyGCY
🔗 API: https://t.co/EOZkbOwCN4

[BitRouter @BitRouterAI]

AGI

[Claude @claudeai]

Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use.

Its capabilities exceed those of any model we’ve ever made generally available.

[BitRouter @BitRouterAI]

someones's wallet gonna be empty, literally

[Polymarket Money @PolymarketMoney]

JUST IN: Anthropic will reportedly release its new AI model “Mythos” tomorrow.

[AshutoshShrivastava @ai_for_success]

Kimi is so back 🔥

Kimi AI has released Kimi Work, a local AI desktop agent built to automate tasks by running parallel agent swarms.

TLDR

- Runs up to 300 AI agents in parallel on your local machine
- Uses WebBridge extension to navigate, search, click, and type in your browser
- Includes native financial data tool calls for Yahoo Finance, World Bank, and Binance
- Features a memory system that logs preferences, past decisions, and context
- Delivers production-ready outputs directly to your desktop in PPTX, Word, PDF, and Excel
- Available for macOS Apple Silicon and Windows

[Xiaomi @Xiaomi]

1000 tokens/s, from a 1 trillion parameter model, on regular GPUs.
Not a lab demo — this is MiMo-V2.5-Pro-UltraSpeed. Speed changes everything.

[Brian Armstrong @brian_armstrong]

Good take

My guess is
- demand for intelligence is near infinite
- but 80% of workloads will be running on 99% cheaper models within 12-18 months
- 20% of workloads will still run on latest gen models where IQ maxing is important (scientific breakthroughs, higher level ochestrator agents?)
- rough analogy might be what % of macbooks or gaming PCs sold have the maxed out specs for CPU/GPU, prices are falling much faster than Moore's law here though
- this leads me to think the limiting factor will be energy and compute, not better models

At Coinbase we're working hard on routing prompts to cheaper models where appropriate, and in some cases have been able to keep costs roughly flat, while token usage continues to grow exponentially.

[BitRouter @BitRouterAI]

open router is not open, we are

[Shaw (spirit/acc) @shawmakesmagic]

OpenRouter is cool

But margin is everything

So we switched to BitRouter, which is free and open source

https://t.co/r3ORONEkQx

[Shaw (spirit/acc) @shawmakesmagic]

OpenRouter is cool

But margin is everything

So we switched to BitRouter, which is free and open source

https://t.co/r3ORONEkQx

[BitRouter @BitRouterAI]

GitHub stars doubles in one day is a great start what what's coming up next

[BitRouter @BitRouterAI]

we built BitRouter as a single binary.

drop it in front of your agent harness and it just works.

no daemon. no sidecar. no yaml sprawl. no infra team required.

the whole point was: if a solo dev can't run it in 60 seconds, we failed.

that constraint shaped almost every product decision we've made.

[BitRouter @BitRouterAI]

@agoraxyz @colosseum @verityxyz @AstariaTrade Glad we made it 🫡

[HackQuest @HackQuest_]

Introducing @BitRouterAI — trustless AI with a permissionless API.

🤝 One unified API for models, tools, and agents
🔓 Permissionless, agent-native infrastructure
⛓️ On-chain, pay-per-inference settlement

Enabling AI agents to operate openly, without gatekeepers.

[BitRouter @BitRouterAI]

Don’t Python🐍 , Rust 🦀

[Andrej Karpathy @karpathy]

Software horror: litellm PyPI supply chain attack.

Simple `pip install litellm` was enough to exfiltrate SSH keys, AWS/GCP/Azure creds, Kubernetes configs, git credentials, env vars (all your API keys), shell history, crypto wallets, SSL private keys, CI/CD secrets, database passwords.

LiteLLM itself has 97 million downloads per month which is already terrible, but much worse, the contagion spreads to any project that depends on litellm. For example, if you did `pip install dspy` (which depended on litellm>=1.64.0), you'd also be pwnd. Same for any other large project that depended on litellm.

Afaict the poisoned version was up for only less than ~1 hour. The attack had a bug which led to its discovery - Callum McMahon was using an MCP plugin inside Cursor that pulled in litellm as a transitive dependency. When litellm 1.82.8 installed, their machine ran out of RAM and crashed. So if the attacker didn't vibe code this attack it could have been undetected for many days or weeks.

Supply chain attacks like this are basically the scariest thing imaginable in modern software. Every time you install any depedency you could be pulling in a poisoned package anywhere deep inside its entire depedency tree. This is especially risky with large projects that might have lots and lots of dependencies. The credentials that do get stolen in each attack can then be used to take over more accounts and compromise more packages.

Classical software engineering would have you believe that dependencies are good (we're building pyramids from bricks), but imo this has to be re-evaluated, and it's why I've been so growingly averse to them, preferring to use LLMs to "yoink" functionality when it's simple enough and possible.