Huddlz - $HUDL (@HuddlzToken) header
Huddlz - $HUDL (@HuddlzToken) profile picture

use @AskMaximoAI - delivering the maximum potential of artificial intelligence to users.

Global
Joined September 2019
Photo by Huddlz - $HUDL (@HuddlzToken)Photo by Huddlz - $HUDL (@HuddlzToken)Photo by Huddlz - $HUDL (@HuddlzToken)Photo by Huddlz - $HUDL (@HuddlzToken)Photo by Huddlz - $HUDL (@HuddlzToken)Photo by Huddlz - $HUDL (@HuddlzToken)Photo by Huddlz - $HUDL (@HuddlzToken)Photo by Huddlz - $HUDL (@HuddlzToken)Photo by Huddlz - $HUDL (@HuddlzToken)Photo by Huddlz - $HUDL (@HuddlzToken)
GPT-6 Astra now available on MyTabulon! Chat: mytabulon.com Coding Plan: platform.mytabulon.com/prici…
GPT-6 Astra is now available on MyTabulon. OpenAI's reasoning flagship — 1.05M context window, 128K max output, reasoning from Low to Max. We wired it into workspace chat (Pro and Max tiers) and Coding Plan (Max) at $10/M in, $1/M cached, $50/M out. Push past 272K prompt tokens and it bumps to $20/$2/$75 for the full request. Same OpenAI-compatible endpoint, full native context, no injected prompts or wrappers. Codex, Kilo Code, OpenClaw, Maximo Syntax CLI — it just works. Chat: mytabulon.com Coding Plan: platform.mytabulon.com/prici…
5
688
Huddlz - $HUDL retweeted
GPT-6 Astra is now available on MyTabulon. OpenAI's reasoning flagship — 1.05M context window, 128K max output, reasoning from Low to Max. We wired it into workspace chat (Pro and Max tiers) and Coding Plan (Max) at $10/M in, $1/M cached, $50/M out. Push past 272K prompt tokens and it bumps to $20/$2/$75 for the full request. Same OpenAI-compatible endpoint, full native context, no injected prompts or wrappers. Codex, Kilo Code, OpenClaw, Maximo Syntax CLI — it just works. Chat: mytabulon.com Coding Plan: platform.mytabulon.com/prici…
This is GPT-6 Astra. Anything you can do on a computer, Astra can do for you. Fast.
3
1
6
974
Huddlz - $HUDL retweeted
Software coding is solved. Honestly, you can't write better code than the latest AI models now, so enough with shaming vibe coding. Frontend and backend, the shit is solved. The focus now is recursive self learning for these models so we get to AGI faster, solve incurable diseases, and start making real world AI impact.
Muse Spark 1.3 is rolling out today with frontier performance almost too cheap to meter. This is the biggest jump we've made so far on coding and agentic work. Try it in Muse Code and our API. Next up 🍉 and Muse Spark open weights releases coming soon.
2
1
7
395
Maximo Atlas 1.3 is the best financial model we've ever trained yet! Try it: maximoai.co/platform
Maximo Atlas 1.3 is the best financial model we've ever trained yet! It's so good at managing trades and any investment portfolio, and it is fast too, making it the better choice compared to Maximo Pandora 3.8 or 3.8 Pro. Crazy.
2
417
Huddlz - $HUDL retweeted
Maximo Atlas 1.3 is the best financial model we've ever trained yet! It's so good at managing trades and any investment portfolio, and it is fast too, making it the better choice compared to Maximo Pandora 3.8 or 3.8 Pro. Crazy.
2
1
6
340
Huddlz - $HUDL retweeted
Meta announced Muse Spark 1.3 today and it is already live on MyTabulon. - Workspace chat: Plus, Pro, and Max workspaces get it as the new current Meta model, with 1M context, Minimal through XHigh reasoning, and native image, PDF, and MP4 video input. Muse Spark 1.2 moves to Legacy in the selector and every saved preference keeps working. - Coding Plan: Plus tiers and up get muse-spark-1.3 at $1.25/M input, $0.15/M cached, $4.25/M output. There is also muse-spark-1.3-contributor over Meta's direct endpoint at $0.10/$0.002/$0.20 per million, with the published no-ZDR tradeoff disclosed right on the pricing page. 1.2 stays in the catalog beside it, so nothing is taken away. Same OpenAI-compatible endpoint, full native 1M context, no injected prompts or tools. Pick it in Codex, Kilo Code, OpenClaw, or the Maximo Syntax CLI and keep coding. Chat: mytabulon.com Coding Plan: platform.mytabulon.com/prici…
We’re excited to release Muse Spark 1.3 with improved performance on agentic and coding tasks, and a focus on real-world usability. Key capabilities: → Sustains longer-horizon work across multiple workflows in a single thread → More actively collaborates with users: it asks clarifying questions, flags when it's stuck, confirms before consequential actions → Better calibrated on its own limits instead of hallucinating outcomes → ~20% fewer tool calls and ~25% fewer tokens vs. Muse Spark 1.2 in internal comparisons
1
2
5
422
Huddlz - $HUDL retweeted
When we shipped Atlas 1.2, the benchmark story was about a model refusing to quit after the first answer. That positioned us well, but there was a ceiling: people still switched models depending on the task. I have been running 1.3 through our own internal workflows for a while now, and the switch point mostly disappeared. Terminal sessions stay coherent. Tool chains finish. BrowseComp and DeepSWE moved double digits from 1.2. Now the part I actually care about. Cached input is $0.02 per 1M tokens. Every agent workload I know re-reads the same context dozens of times per session, and that cached read is where bills quietly explode. We priced it like it should be nearly free and put the real price on output, where value is created. Frontier capability that rations itself is a demo. This one is built to run all day.
We're introducing Maximo Atlas 1.3. A full-stack frontier agent built for terminal, code, and web work at max reasoning effort. Now live on Maximo AI API Platform at $0.20 input, $0.02 cached input, and $0.50 output per 1M tokens. maximoai.co/platform
3
8
277
We're introducing Maximo Atlas 1.3. A full-stack frontier agent built for terminal, code, and web work at max reasoning effort. Now live on Maximo AI API Platform at $0.20 input, $0.02 cached input, and $0.50 output per 1M tokens. maximoai.co/platform
We're introducing Maximo Atlas 1.3. A full-stack frontier agent built for terminal, code, and web work at max reasoning effort. Now live on Maximo AI API Platform at $0.20 input, $0.02 cached input, and $0.50 output per 1M tokens. maximoai.co/platform
2
570
Huddlz - $HUDL retweeted
We're introducing Maximo Atlas 1.3. A full-stack frontier agent built for terminal, code, and web work at max reasoning effort. Now live on Maximo AI API Platform at $0.20 input, $0.02 cached input, and $0.50 output per 1M tokens. maximoai.co/platform
1
3
2
5
828
Huddlz - $HUDL retweeted
Atlas 1.3 scores 84.4% on Terminal-Bench 2.1, up from 82.7% for Atlas 1.2. Deeper: 39.0% on Terminal-Bench 4.0, 64.0% on DeepSWE 1.1, and 1774 Elo on GDPval-AA v2.
1
1
4
61
Huddlz - $HUDL retweeted
Cached input with Atlas 1.3 costs 90% less than uncached input. For long agent sessions that re-read the same context, that is where the bill actually goes. Full benchmark matrix, methodology, and pricing in our launch post: maximoai.medium.com/maximo-a…
1
4
51
So good!
Replying to @OpulentByte
Now see what Maximo Atlas 1.3 did, a very big leap!
2
577
Huddlz - $HUDL retweeted
Holy shit! Maximo Atlas 1.3 model is a GOAT! Unbelievable! This first video is a motion video done by Maximo Atlas 1.2 model. Just terrible at complex UI/UX.
1
1
4
130
Huddlz - $HUDL retweeted
This is what Maximo Atlas 1.2 did.
1
1
4
100
Huddlz - $HUDL retweeted
Now see what Maximo Atlas 1.3 did, a very big leap!
1
1
1
3
519
Huddlz - $HUDL retweeted
All videos were done using Maximo Syntax Desktop. Maximo Syntax Desktop is not just an app for coding agent, it is a tool for work! Do anything! Dm me for early access to the desktop app.
1
2
43
Huddlz - $HUDL retweeted
Update: Maximo Atlas 1.3 and 1.4 are back online. Thanks for your patience.
Heads up: Maximo Atlas 1.3 and 1.4 Experimental Preview are temporarily down. We are actively working on bringing them back online. Other models remain available. We'll update here when they're live again.
2
2
4
182
Huddlz - $HUDL retweeted
Heads up: Maximo Atlas 1.3 and 1.4 Experimental Preview are temporarily down. We are actively working on bringing them back online. Other models remain available. We'll update here when they're live again.
2
1
4
275
Huddlz - $HUDL retweeted
Most agent tooling speaks the Responses API now, not just Chat Completions. We did not want anyone maintaining two stacks just to use our models. So we built /v1/responses directly on the same infrastructure as chat completions. Same model registry, same auth, same USD billing. Nothing new to configure, nothing to migrate. The part I find underrated: if your tool already targets the spec, using Maximo models becomes a base URL swap.
/v1/responses is now live. Streaming, stored responses, previous_response_id chaining and tool calling. Same auth and USD billing as /v1/chat/completions. Already on the Responses API spec? Point it at our base URL with your key and you're done. maximoai.co/platform
1
5
155
So we built /v1/responses directly on the same infrastructure as chat completions. Same model registry, same auth, same USD billing. Nothing new to configure, nothing to migrate.
/v1/responses is now live. Streaming, stored responses, previous_response_id chaining and tool calling. Same auth and USD billing as /v1/chat/completions. Already on the Responses API spec? Point it at our base URL with your key and you're done. maximoai.co/platform
4
653