Building engineering intelligence @ Entelligence.AI
- Tweets 882
- Following 9
- Followers 4,990
- Likes 3,802
Pinned Tweet
We’re hiring an SDR at Entelligence AI.
You’ll sell to VPs of Engineering, CTOs and platform leaders at companies running coding agents at scale.
Looking for someone with 1-3 years of SDR experience, or a technical background moving into sales, who understands code review, token spend and model routing.
You’ll work directly with the founding team and CEO.
Apply link below 👇
1
2
19
3,716
Entelligence AI retweeted
Had a great time at WeAreDevelopers Congress!
Dug into what it takes to measure coding agent ROI at JPMorgan scale with their Payments team
How Entelligence.AI helps:
- Model Router: route each task to the best-performing model
- Agent Insights: share the skills behind successful sessions
- Pinpoint where agents loop and burn tokens
- World Model: agents learn from every past error and incident
1
3
18
780
HOW MUCH AI ARE YOUR DEVELOPERS ACTUALLY USING?
one of our developers used 4.1B tokens and $2.8K+ in AI spend, across Claude Code, Cursor, Codex and multiple frontier models
92.2% of the token usage came from Claude Opus 5
this is what it looks like in Agent Insights ↓
1
1
7
292
your coding agents are probably using way more skills and sub-agents than you realize.
Agent Insights now lets you see exactly which skills your developers are using across their coding sessions, including namespaced/plugin skills and sub-agent usage
16 skills used by this developer alone ↓
1
7
285
Opus 5.5, GPT-6 Sol, Grok 4.7, Qwen…
your coding agent probably doesn’t need the same model for every turn, some tasks need frontier-level intelligence but MOSTLY don't.
that’s what Model Router does, routes each turn to the model that makes sense and escalates when it actually needs more intelligence.
We ran 810 coding-agent turns through it and saw 32.7% lower cost vs direct model calls.
Results ↓
1
1
4
423
try it out here-
entelligence.ai/model-router
1
1
87
Our Model Router now supports the latest flagship models !!!
Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, Grok 4.7 and more.
If you’re already using flagship models, there’s never been a better time to start optimizing your costs.
1
2
8
266
Spending $1K+ a month on AI? You should know where every dollar is going.
Agent Insights lets you:
1. track spend across your entire engineering team
2. see which agents, sessions & developers drive the most usage
3. break down token, model & harness usage
4. uncover patterns and areas worth optimizing
Here’s what it looks like ↓
2
4
8
468
We’re hiring an SDR at Entelligence AI.
You’ll sell to VPs of Engineering, CTOs and platform leaders at companies running coding agents at scale.
Looking for someone with 1-3 years of SDR experience, or a technical background moving into sales, who understands code review, token spend and model routing.
You’ll work directly with the founding team and CEO.
Apply link below 👇
1
2
19
3,716
apply here - linkedin.com/jobs/view/44661…
139
Entelligence AI retweeted
Entelligence will be at WeAreDevelopers World Congress North America, Sept 23–25! 🚀
Building with AI coding agents? Let’s talk smarter model routing, less token waste, and better code.
Made with AI
1
3
6
606
Opus 5.5 just dropped at $20/M output tokens
meanwhile, your agent could be routing simple tasks to cheaper models and only escalating when it actually needs Opus-level intelligence.
that’s what Model Router does.
Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family.
It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.
1
1
7
809
check it out here - entelligence.ai/model-router
1
2
103
we tested a simple code review routing strategy:
run GPT-5.6 Luna on every PR, then bring in GPT-5.6 Sol when Luna finds something on 50 real PRs:
• 120 verified bugs vs 108 with Sol alone
• $4.13 total cost vs $4.15
• Luna’s findings were a better escalation signal than diff size or file paths
2
1
6
356
full benchmark ↓
entelligence.ai/blogs/cheap-…
1
2
91
another model joins the race!
meanwhile, we’re making sure your agents don’t have to pick one model and stick with it
Model Router automatically routes each request to the right model for the job.
MiMo-V2.6-Pro debuts as the top open weights model on the Artificial Analysis Intelligence Index (46). At $0.13 per Intelligence Index task, it lands on the Intelligence vs. Cost per Task Pareto frontier
@Xiaomi has just released MiMo-V2.6-Pro, an open weights model with major advances in intelligence over its predecessor, MiMo-V2.5-Pro (Intelligence Index: 26). Despite the improvement, it retains the same attractive pricing at $0.435 per 1M input tokens (with a 99% cache-hit discount) and $0.87 per 1M output tokens. This makes MiMo-V2.6-Pro one of the most cost-efficient models to deploy.
MiMo-V2.6-Pro is an MoE model with 1.02T total parameters and 42B active parameters. Stay tuned for additional analysis of the model.
Check out MiMo-V2.6-Pro full benchmarking breakdown here: artificialanalysis.ai
2
1
9
924
Entelligence AI retweeted
moving towards a future where your coding agents will Literally never make the same errors again
super exciting to work on this at an enterprise scale.
the data we're gathering about coding agent failure modes is probably the most interesting dataset I've worked with to date
7
3
20
1,183
Entelligence AI retweeted
Kicking off LDX3
If you’re attending come find the @EntelligenceAI team floor 5!
Lots of great discussions
3
25
1,178
Entelligence AI retweeted
Astra and Fable came out and there are still as many issues in PRs as before
Coding agents need production context to learn and prevent historical issues
Launching the @EntelligenceAI World model - stores every historical session, PR review comment, and incident giving your coding agents production context to prevent repeat errors.
3
3
13
1,012
woke up to this from @EntelligenceAI:
- 6k+ visitors on the latest blog in 12 hours
- ~100k views across Reddit
- 100+ comments on Hacker News
i genuinely don't know what happened while i was asleep 😭
7
4
46
1,636
can a $1.20/M token model actually compete with the best model on the internet right now?
we put GPT-5.6 Luna against GPT-6 Astra on 50 real PRs and Astra found 92 verified bugs vs 69 for Luna
Luna was able to catch 75% of the bugs for just 3.6% of the cost
honestly didn’t expect the gap to be this small
3
3
9
779
full eval + bug breakdown here:
entelligence.ai/blogs/gpt-5.…
1
2
360