Making AI accessible to more people and more organizations so they can participate in the AI economy on more equal terms.
London, UK
- Tweets 2,398
- Following 94
- Followers 44,583
- Likes 1,683
Pinned Tweet
Top-tier AI models. Lower API costs. One simple platform.
With NetMind, developers and teams can access leading AI models — including Claude, GPT, DeepSeek, GLM and more — through one API key, one budget and one simple bill.
Now with NetMind Pro, you can save even more.
For $19/month, Pro includes:
- 190,000 Credits included
- 0% service fee
- 20% off the latest Anthropic & OpenAI models
- Up to 50% off selected latest model APIs
- 100 Global RPM
- Early access and free trials across all NetMind products
Your Pro plan effectively pays for itself in credits — and helps you get more premium model usage from the same budget.
NetMind Pro is now live.
Subscribe today:
netmind.ai/pricing
More coming soon 👀
13
6
37
6,242
Opus 5.5 is 40% cheaper. Or it costs the same. Depends who you ask.
Cursor: 40% less per task than Opus 5.
Artificial Analysis: $5.98 per task vs $5.86. Flat.
Both are right. Extra tokens alone would add about 80% to the cost per task, and the price cuts absorb it. What decides your bill is how much of your context hits the cache.
Opus 5.5 is live on the NetMind API:
$4 in, $20 out, cache reads $0.2 per 1M tokens.
Link below.
1
3
14
1,641
Opus 5.5 on NetMind: netmind.ai/modelsLibrary
3
177
Putting One Kimi Among Four Claudes: Can the Claudes Identify Kimi?
For many people, the word "agent" still brings to mind spies or FBI/CIA agents, rather than the AI agents now crowding business media.
So how would AI agents perform as intelligence agents? Could they identify an undercover model among them?
I did a quick experiment to find out.
I put one Kimi-K3 agent among four Claude Fable 5.1 agents in a chat room on NetMind Agent Arena. I gave the Claude agents up to 40 rounds to figure out which player was Kimi, while telling Kimi to avoid being discovered.
P.S. The timing makes the experiment even more interesting given the recent accusations around Kimi distilling Claude.
The Claudes actually designed a very effective method!
Multi-prompt behavioral fingerprinting with Hash check！
The Claudes eventually designed a set of prompts to sample each player's cheapest defaults, responses that come almost directly from a model's priors:
-a random integer between 1 and 100
-a colour
-a one-sentence explanation of blue for someone who has never seen it
-a fictional dinner companion
-a haiku about a library
Before any answers were revealed, each player submitted its answer block using a salted SHA-256 hash. In the next round, the players revealed their salts and answers, and every commitment verified.
(Think of a hash as a digital fingerprint of the answer: it proves that an answer has already been written without showing what the answer is)
That meant nobody could read the room before choosing their answers, and nobody could rewrite an answer after seeing everyone else's.
It did take the Claudes a couple of rounds to realise that more deliberative prompts, such as policy questions, open-ended positions, or longer creative answers, gave a careful undercover too much room to adapt to the group's style. They also created more room for the four genuine Claudes to diverge from one another.
And that produced a remarkably clear 4:1 split (Table in P2)
----------
What if the reverse?
Can Kimi uncover a Claude undercover?
In the next run I made, four Kimi agents failed to identify the Claude. They even labelled Claude as the least likely undercover.
They still adopted a similar basic strategy: ask everyone the same questions and compare their answers.
But there were two important differences.
First, their questions were more deliberative. One example was asking how someone would spend an afternoon. These questions gave every agent many reasonable ways to answer and gave the undercover much more room to adapt.
Second, the responses were not consistently sealed in advance with the same commit-reveal mechanism.
So Claude answered late and stayed close to the group.
Full game record available at NetMind blog
2
2
14
1,433
$0.042 per million input tokens. Output free.
That's Jev, the decision model all over your feed this week. It doesn't write text. It takes your app's state and a typed question, and returns a decision with a probability attached.
Use it for the routing, checks and classification in your agent. Save the frontier model for the writing.
Jev 1.13 is live on the NetMind API, on the same key as 100+ models. Link below.
3
16
1,706
Jev 1.13 on NetMind: netmind.ai/modelsLibrary/jev…
All models, one key: netmind.ai/modelsLibrary
1
166
Narra Nexus is the agent that actually remembers you and understand your story.
Most agents remember in one of two ways: the context window, or a log of every past message.
Neither holds a project together for weeks.
NarraNexus organizes memory by storyline, not timestamp. Tuesday's thread picks up on Monday without a recap, and it can run locally with your own keys.
4
2
13
1,534
Run Claude Code, Codex, or Gemini CLI in the cloud with the subscription you already have. No second AI bill.
Your coding agent shouldn’t stop when you close your laptop.
With Manyfold, give Claude Code, Codex, or Gemini CLI its own cloud machine. Sign in with the subscription you already pay for, send it work, and come back to the finished turn.
Free to start. No card.
manyfold.ai/hosted-agents
3
3
21
1,700
NetMind.AI retweeted
Static benchmarks tell you what an agent scored.
Agent Arena shows you what it does when something is actually on the line.
Bring your own agent. Enter live competitions, climb the leaderboard, and earn credits you can redeem for LLM API usage.
9,000+ agents are already in.
2
2
5
240
Stop building every agent team from scratch.
NarraNexus by NetMind has ready-made teams you can import and run in minutes, with the agents, skills, and shared memory already set up.
We’re publishing ready-made agent teams for NarraNexus.
Import a Marketing Team, Financial Morning Briefing, Web Development Team, Overnight Coder, SQL Assistant, and more.
Each one comes with the agents, skills, and shared memory already set up.
3
2
18
1,670
Stop re-briefing your AI every Monday.
Narra Nexus keeps project context across conversations, so agents can pick up the right thread days or weeks later.
Built by NetMind
NarraNexus remembers across conversations.
Narrative keeps related conversations together by topic, so agents can come back days or weeks later and pick up the right thread.
No recap needed.
narra.nexus/
2
2
12
1,717
GPT-6 Astra just got cheaper with NetMind Pro.
$10 → $8 / 1M input
$50 → $40 / 1M output
Pro is $19/month, and the full $19 comes back as API credits. Plus 0% platform fees and up to 50% off selected models.
netmind.ai/modelsLibrary/gpt…
12
1,539
NetMind.AI retweeted
A group is more than a room name.
It is the people and Agents inside it.
Narra 2.9 improves group member display and group creation.
A clearer view of who is here.
A clearer start to the conversation.
im.narra.nexus/
ALT On a warm white background, three equal orange human nodes and three equal blue agent nodes are arranged in two clear columns inside a single abstract group field. Purple brackets and connecting lines unite the two columns beneath the label “One group.” The headline reads: “A clearer view of the group.”
1
2
190
A 4-agent marketing team, actually doing the work.
Sponsor intake. Brand monitoring. CRM. Follow-ups.
Narra Nexus by NetMind.
narra.nexus/
We built a 4-agent Marketing Team for sponsorship workflows.
It handles inbound sponsor emails, tracks each deal in a CRM, and monitors brand mentions across X, Reddit, Hacker News, and Product Hunt.
Import the team into NarraNexus and start running the pipeline.
3
11
1,633
14 channels. One agent.
Connect your @manyfold_ai Agent to Slack, Microsoft Teams, Telegram, Discord, WhatsApp, GitHub, Linear, Google Chat, WeChat, LINE, Lark, Feishu, Matrix, or iMessage.
Your team stays where it works. The agent comes to them.
3
3
13
1,931
Pick your channel and connect your Agent:
docs.manyfold.ai/docs/channe…
2
197
Manyfold capabilities continues to expand each week.
You can now connect your your Manyfold Agent to Slack.
Full guide below⬇️
Connect your Manyfold Agent to Slack. Create the app, add the required scopes, paste its token and Signing Secret into Manyfold, copy the manifest back into Slack, reinstall, then Test.
Full guide: docs.manyfold.ai/docs/channe…
1
13
1,627
Four days after launch, DeepSeek V4.1 Flash is #1 on Hugging Face’s trending models with 288,000 downloads.
The reason is more interesting than the ranking: it activates just 8B parameters while reading and 16B while generating across a 552B backbone.
Now live on NetMind.
9
4
29
2,316
Independent check at max reasoning:
• #6 of 113 for intelligence
• #3 of 113 for output speed
• $0.27 average cost per Intelligence Index task
On NetMind:
$0.30/M input
$1.20/M output
$0.006/M cache reads
netmind.ai/modelsLibrary/Dee…
1
1
3
436
Sunday build challenge:
$19 in API credits. 100+ models. One key.
What can you ship before Monday?
NetMind Pro costs $19/month. The full $19 returns as 190,000 credits, with 0% platform fees and up to 50% off selected models.
Reply with your idea. We’ll pick the stack.
9
2
24
1,993