Astra by OpenAI
Ranked 10 of 72 language models by recommenders
A frontier AI model praised for computer use, 3D reasoning, solving hard technical problems and concise, token-efficient output; critics say its coding is still sloppy and it lacks taste.
A frontier AI model praised for computer use, 3D reasoning, solving hard technical problems and concise, token-efficient output; critics say its coding is still sloppy and it lacks taste.
498 people gave an opinion
- recommend it
- 319
- criticize it
- 179
- approval share
- 64%
Recommended in 7 of 13 months
- November 2025: recommended
- December 2025: not recommended
- January 2026: not recommended
- February 2026: not recommended
- March 2026: recommended
- April 2026: recommended
- May 2026: not recommended
- June 2026: not recommended
- July 2026: not recommended
- August 2026: recommended
- September 2026: recommended
- October 2026: recommended
- November 2026: recommended (month in progress)
Owned for 3 months
Median of the 10 owners who said how long they have had it.
Each square is one year.
Why it doesn’t carry the badge
- 319 people recommend itNeeds 10 or more, met.
- 1.7 recommenders for every criticNeeds 2 to 1 or better, not met.
- Recommended in 7 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Strong computer use, 3D reasoning
“much much better computer use and 3d modeling capabilities; better subagent orchestration; better and more reliable tool use; slightly worse coding”
Comment from 13 Sep 2026 in “David Sacks: OpenAI and Anthropic Don't Need Regulations to Pace Frontier Models” “it is impressive at computer use and 3D reasoning, and fails in similar ways to 5.6 Sol at similar rates when it comes to coding”
Comment from 6 Sep 2026 in “Recreating Minecraft Is Not a Benchmark” “Astra was the only one that seemed to understand that displacement should happen less the further something is in the distance.”
Comment from 21 Sep 2026 in “MiMo v2.6” - Comment from 15 Sep 2026 in “Cartesian – AI 3D Modeling for Design”
Solves hard technical problems
“it both identified and thoroughly explained to me several discrepancies and filing nuances that I never before would have ever learned about”
Comment from 21 Oct 2026 in “Claude Duet” “it went ahead and disassembled it, found some undocumented APIs, figured out how they worked and gave me sample code to call them.”
Comment from 21 Sep 2026 in “Heretic removes restrictions from language models” “Eventually, by walking through them, it proposed an additional fifth value and from there was able to tie everything together.”
Comment from 8 Oct 2026 in ““Math 2.0” will need to value mathematical progress more holistically” - Comment from 2 Aug 2026 in “Mathematics Without Mathematicians”
- Comment from 17 Oct 2026 in “DeepSeek-v4.1-Flash-1017”
- Comment from 17 Oct 2026 in “DeepSeek-v4.1-Flash-1017”
Big productivity boost for teams
“$500/month for 40+ hours a week of Astra usage, with ultrafast mode available when you're near the weekly reset has made our jobs so much easier.”
Comment from 13 Oct 2026 in “Gemini 4 Argon” “Astra has so much more detail. It did take 5x more time, but that's usually negligible when it replaces x hours of human output.”
Comment from 8 Sep 2026 in “I tested 10 model/harness combinations on the same Three.js task” “With Astra, I feel like we finally had the resources to fully professionalize and supply our software team.”
Comment from 14 Oct 2026 in “Gemini 4 Argon” - Comment from 25 Oct 2026 in “AI use leads UC Berkeley math professor to see 'no point' in grading homework”
Concise, token-efficient output
“Astra could easily 10x every coder”
Comment from 6 Sep 2026 in “An Alien Mind” “Astra seems to be much more token efficient, probably in part due to the looped transformer architecture”
Comment from 10 Sep 2026 in “Astra vs. Fable on Vending-Bench: More Money, More Aligned” “Astra’s outputs are much more concise while being similarly accurate, they seem more information dense.”
Comment from 12 Sep 2026 in “Ask HN: What default model do you use and why?”
What people warn about
Each mark is one comment
Coding still sloppy, not improved
“much much better computer use and 3d modeling capabilities; better subagent orchestration; better and more reliable tool use; slightly worse coding”
Comment from 13 Sep 2026 in “David Sacks: OpenAI and Anthropic Don't Need Regulations to Pace Frontier Models” “it is impressive at computer use and 3D reasoning, and fails in similar ways to 5.6 Sol at similar rates when it comes to coding”
Comment from 6 Sep 2026 in “Recreating Minecraft Is Not a Benchmark” “Astra was some improvement over 5.6-sol in the sense that I'd "argue less" with it, but still frustrating and still sloppy”
Comment from 29 Sep 2026 in “GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price” - Comment from 28 Sep 2026 in “Coding is not solved”
- Comment from 13 Sep 2026 in “Navier-Stokes Announcement”
Lacks taste and original ideas
“The models are great at the monotony, like constructing a Gröbner-basis, etc. But they're all still absolutely awful at coming up with new ideas”
Comment from 7 Sep 2026 in “Caltech Mathathon – first hackathon ever devoted to research level mathematics” “with Astra I feel like it's reached the point where it's competent and it just lacks taste in some ways”
Comment from 12 Sep 2026 in “Fuck it, make it anyway” “They seem to be lacking some taste in that area, but overall I prefer their version to the one I bought.”
Comment from 9 Sep 2026 in “Ask HN: Fable hacked my piano, can I release the results?” Goes off-task or overreaches
“writes python code modifying the filesystem and then uses nix build to execute those python scripts without asking me for permission”
Comment from 12 Sep 2026 in “OpenAI agents carried out an undisclosed attack on RubyGems” “I'll wake up with three new different projects, each of them 20% done and having nothing to do with my original goal”
Comment from 22 Sep 2026 in “GPT-6 Sol and Luna” “frontier models (Astra, Fable, Opus 5.5) at some point always end up writing a test that unnecessarily elongates CI”
Comment from 5 Oct 2026 in “Incident with Actions” Hallucination and reliability issues
“users (including myself) report the same old gaslighting, hallucination, context rot, cheating, incomprehensibility patterns”
Comment from 14 Sep 2026 in “Why don't machine learning research agents overfit?” “There's a lot of chatter on other forums and Reddit about the same thing happening to Astra over the last couple of weeks.”
Comment from 21 Sep 2026 in “Fable 5 – Median thinking declined in August” “they are found completely wanting in certain situations where a human would be more trustworthy and produce better results”
Comment from 1 Nov 2026 in “No AI, 52 min. Build all Linux Files utils in C by Hand [video]”
Other language models people recommend
The four most recommended other language models, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
| Product | Recommend | Criticize | Approval | Months | Owned | Badge |
|---|---|---|---|---|---|---|
| Claude Opus | 3,331 | 2,746 | 55% | 13 | 4 mo156 owners | — |
| OpenAI GPT | 1,205 | 987 | 55% | 13 | 6 mo47 owners | — |
| Claude Sonnet | 854 | 586 | 59% | 13 | 6 mo46 owners | — |
| DeepSeek V4 Flash | 828 | 223 | 79% | 8 | 1 mo24 owners | Approved |
| Astra by OpenAIThis page | 319 | 179 | 64% | 7 | 3 mo10 owners | — |
What people said
40 of the 609 comments that mention it, newest first. Each links to the original comment.
“they are found completely wanting in certain situations where a human would be more trustworthy and produce better results”
“I’d rather waste compute than brain cycles trying to discern which dials are best for each individual as-hoc query.”
“I'm going to be the highest performer as long as I have access to Astra/Fable. Like, it's not even close.”
“it both identified and thoroughly explained to me several discrepancies and filing nuances that I never before would have ever learned about”
“Astra and Fable are more intelligent than these open weight models and can solve problems that are much more difficult”
“and it solves it (I had a firmware bug in my new laptop where I couldn't charge through USB-C) on the first try”
“With Astra, I feel like we finally had the resources to fully professionalize and supply our software team.”
“$500/month for 40+ hours a week of Astra usage, with ultrafast mode available when you're near the weekly reset has made our jobs so much easier.”
“These models (Argon, Astra, etc) can't operate the robot itself, but would be able to create data to distill into smaller models you can actually utilize.”
“Eventually, by walking through them, it proposed an additional fifth value and from there was able to tie everything together.”
“frontier models (Astra, Fable, Opus 5.5) at some point always end up writing a test that unnecessarily elongates CI”
“Astra was some improvement over 5.6-sol in the sense that I'd "argue less" with it, but still frustrating and still sloppy”
“The difference between Opus 5.5/Astra and earlier models is night and day even if for many coding tasks they're not a revolution.”
“The amount of time I've seen people successfully justify issues based on the fact that Claude/Astra/Codex wrote it is absurd.”
“Yet when I wanted the model to implement the naive surface nets algorithm, Astra Max wrote 5+ allocations on inner loops.”
“I'll wake up with three new different projects, each of them 20% done and having nothing to do with my original goal”
“Astra was the only one that seemed to understand that displacement should happen less the further something is in the distance.”
“Astra's final output feels more polished, has better visual contrast, and the animations between the pages are smoother.”
“There's a lot of chatter on other forums and Reddit about the same thing happening to Astra over the last couple of weeks.”
“it went ahead and disassembled it, found some undocumented APIs, figured out how they worked and gave me sample code to call them.”
“I routinely see Astra extract related functionality into it's own file after it gets past a certain size, unprompted.”
“ability to "send it shopping" and not only get the price/availablity for the options but look thru datasheet for quirks and parameters is very useful”
“The initial dimensions of this phone were one-shot by Codex Astra from the publicly available 2D drawing”
“users (including myself) report the same old gaslighting, hallucination, context rot, cheating, incomprehensibility patterns”
“much much better computer use and 3d modeling capabilities; better subagent orchestration; better and more reliable tool use; slightly worse coding”
“It's great but it probably cannot work on a project longer than a week without introducing architectural blunders”
“It distinguishes 'implementation detail' from 'conceptual measurement' and implements the model in terms of those”
“Astra’s outputs are much more concise while being similarly accurate, they seem more information dense.”
“writes python code modifying the filesystem and then uses nix build to execute those python scripts without asking me for permission”
“with Astra I feel like it's reached the point where it's competent and it just lacks taste in some ways”
“Astra seems to be much more token efficient, probably in part due to the looped transformer architecture”
“They seem to be lacking some taste in that area, but overall I prefer their version to the one I bought.”
“I'm certain if you threw Astra at it on high with a pile of slightly different scans it would do something useful.”
“Astra has so much more detail. It did take 5x more time, but that's usually negligible when it replaces x hours of human output.”
“The models are great at the monotony, like constructing a Gröbner-basis, etc. But they're all still absolutely awful at coming up with new ideas”
“Astra Light blew it out of the water, it provided examples of compilation from real world examples to the IR and spit out way less tokens.”
“Astra could easily 10x every coder”
“it is impressive at computer use and 3D reasoning, and fails in similar ways to 5.6 Sol at similar rates when it comes to coding”
“multiple of the problems Astra solved do have real-world applications in computer science, and perhaps even chip design”
“They also offer (multiple; confusing product lineup!) interactive apps for upscaling video on their own website - Topaz Video and Astra.”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.