GPT-5.6 Sol

What hackers pick instead

• Claude OpusRecommended as a top coding model that fixes bugs and ships features with fewer mistakes than earlier LLMs, though critics cite weak architecture choices, usage limits and confident errors.3331recommend
• OpenAI GPTRecommended for strong coding, debugging and research help with good reasoning and token efficiency, though critics cite hallucinations, overcomplication, verbose output and high cost.1205recommend
• Claude SonnetA capable coding model that has improved noticeably with each version and is faster and cheaper than Opus, though it needs careful supervision and usage limits bite.854recommend
• DeepSeek V4 FlashA very cheap, fast model that people find close to frontier quality for coding and everyday tasks; critics say it errs more, loops, and needs steering.828recommend
• FableRecommended as a top coding model for hard, long-running tasks and large refactors, though critics cite over-engineered code, costly usage limits and safety guards that switch models mid-task.823recommend

Every quote

The 40 most detailed of 1615 comments. Each links to the original.

1. pairing Sol Medium for architecting with Luna High for coding does kinda make sense. But beware that architecting can be very read-heavy, and Sol is a bit read-pricey compared to Terra.

   Mixed · Jul 2026 · in “Advancing the price-performance frontier with GPT‑5.6”

2. I'm often amazed by the gap between how well a frontier model (GPT 5.6 Sol in my case) can comprehend code and investigate a bug, and how badly it writes code and documentation

   Mixed · Jul 2026 · in “OpenJDK Interim Policy on Generative AI”

3. GPT-5.6 Sol dominated the practical details that decide real-world usefulness: tighter instruction-following, cleaner formatting, and fewer correctness slips.

   Recommends · Jul 2026 · in “Head to head: GLM 5.2 vs. OpenAI: GPT-5.6 Sol”

4. I've learned how to deal with and steer Sol medium quite efficiently, but at the same time I realize it's so slow for the small tasks it can do well

   Mixed · Aug 2026 · in “Accelerating GPT-5.6 Sol Ultrafast”

5. I do prefer GPT-5.6 Sol and GPT-6 Astra for producing code that I personally feel is more tasteful, with comments that are less stupid and redundant

   Mixed · Sep 2026 · in “Programming is Art”

6. Cutting the price by 50% while pushing a model this capable is exactly the kind of move that makes advanced AI feel genuinely accessible.

   Recommends · Aug 2026 · in “GPT-5.6 Sol Pricing Cut by 50%”

7. I can tell 5.6 Sol to do something like "optimize this entire subsystem to have no ongoing memory allocations" and go work on other code

   Recommends · Jul 2026 · in “I Think You Might Be Fooling Yourself with AI”

8. Even GPT-5.6 Sol tends to be easier on the eyes for me (though the stereotype of it overengineering and no common sense are still true)

   Mixed · Aug 2026 · in “How I use LLMs to learn complex topics”

9. It's been very successful at frontier math tasks - a bunch of the Erdos questions have been solved by it - more than any other model.

   Recommends · Jul 2026 · in “GPT-5.6 Sol Ultra will be in Codex”

10. GPT-5.6 just finished writing a shim for me in rust that sits between Broadcom's bullshit kernel and a standard Linux user space

    Recommends · Jul 2026 · in “The zero-cost fallacy: open-source software in the agentic era”

11. over the course of ~1 week I was able to review, plan, implement, test and release a significant change to a production system

    Recommends · Aug 2026 · in “GPT-5.6 Sol Pricing Cut by 50%”

12. I’ve interestingly found 5.6 sol to produce much better writing, and it can generally cut to the point much more effectively.

    Recommends · Jul 2026 · in “Claude Opus 5”

13. 5.6 at max effort and fable have done tasks for me that I previously that would be a nearly 6mo project and it took me a week

    Recommends · Sep 2026 · in “GPT-6 Astra”

14. 5.6 Pro is genuinely pretty good at certain kinds of math problems that just require trying out lots and lots of solutions

    Mixed · Jul 2026 · in “GPT-5.6 used a prompt to close a 30-year gap in convex optimization”

15. They all work, they all are "good", they all are both "smart" and commit incredible basic mistakes a fair amount of times.

    Mixed · Sep 2026 · in “Claude Fable 5.1 and Claude Mythos 5.1”

16. 5.6-Sol on xhigh achieves a slightly higher score than Fable 5, but consuming half the tokens and at about 1/3rd the cost

    Recommends · Jul 2026 · in “Fable 5 vs. GPT-5.6 Sol on an NP-Hard Problem: Does /goal help?”

17. $1.20 per 1M output is crazy, but let’s also not forget that OpenAI was one of the first to start increasing model prices

    Mixed · Jul 2026 · in “Advancing the price-performance frontier with GPT‑5.6”

18. GPT 5.6 can be prone to overengineering, but with a little steering and good judgement it generates very nice interfaces

    Mixed · Aug 2026 · in “GPT-5.6 Sol Pricing Cut by 50%”

19. Sol is great at talking to the human and orchestration of agent calls, but it's just too expensive to use everywhere.

    Mixed · Jul 2026 · in “Migrating a production AI agent to GPT-5.6: 2.2x faster, 27% cheaper”

20. sometimes it seems great, other times it makes mistakes so stupid I cannot understand how it ever gets anything right

    Mixed · Aug 2026 · in “Grok 4.6”

21. 1:1 expansion, when I explain what needs to be said and model rewrites it into 1-2 sentences is mostly undetectable.

    Recommends · Sep 2026 · in “The revolt of the reader”

22. Fable and GPT-5.6 chewed through the compiler warnings, dependency hell, and portability bugs largely on their own.

    Recommends · Jul 2026 · in “Ask HN: What Are You Working On? (July 2026)”

23. My results with 5.6 Sol were quite similar to 6, although I haven't tested it that much.

    Recommends · Sep 2026 · in “GPT-6 Sol and Luna”

24. it has persistently written false information when it had the correct information within its context to a degree I've not experienced with other models

    Criticizes · Aug 2026 · in “OpenAI: GPT 5.6 Sol price reduction (until at least Nov 21)”

25. In our previous eval rounds Sol was more expensive than Opus (with its original price), took much longer, and provided worse results.

    Criticizes · Aug 2026 · in “GPT-5.6 Sol Pricing Cut by 50%”

26. I found 6 to burn through my usage incredibly quick, making it somewhat unusable because I wouldn't be able to get anything done.

    Criticizes · Sep 2026 · in “GPT-6 Sol and Luna”

27. I had not hit any guardrails with reverse engineering (as long as its not related to security) with GPT 5.5 or 5.6 Sol, so you should try those.

    Recommends · Aug 2026 · in “Responding to the next frontier of critical cyber capabilities”

28. This is already more intelligent, more imaginative, more disciplined, and far more erudite than the vast majority of individual humans.

    Recommends · Oct 2026 · in “Artificial Insanity”

29. This does look pretty incredible, but don't forget that incredible token thoroughput can only necessarily solve certain bottlenecks.

    Mixed · Aug 2026 · in “Accelerating GPT-5.6 Sol Ultrafast”

30. for me compaction is a signal to restart the session or to tell the LLM to reference existing design docs lest we go off the rails.

    Mixed · Jul 2026 · in “OpenAI reduces Codex Model Context Size from 372k to 272k”

31. I started sending well defined review packages to Luna x-high instead, which is how I have this setup when using Codex (Sol drives

    Recommends · Sep 2026 · in “GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price”

32. Pokemon is no longer a difficult challenge for it, even with additional difficulty restrictions such as hard mode and level caps.

    Recommends · Oct 2026 · in “GPT-5.6 Sol can beat Factorio on peaceful mode”

33. they're clearly changing things behind the scenes a lot. GPT-5.6 for example should not be constantly tweaked without telling us

    Mixed · Jul 2026 · in “Terrence Tao's ChatGPT Conversation about the Jacobian Conjecture Counterexample”

34. I use GLM5.2 and gpt-5.5 and 5.6 and never noticed any quality degradation near the limits of the context or due to compaction.

    Recommends · Jul 2026 · in “OpenAI reduces Codex Model Context Size from 372k to 272k”

35. it blew me away with how "smart" it seemed, how it completely sidestepped failure modes I'd been used to dealing with in 5.5

    Recommends · Jul 2026 · in “I Think You Might Be Fooling Yourself with AI”

36. The price per token increases when >256k though, but it can be useful when compaction at-that-moment would be detrimental.

    Mixed · Aug 2026 · in “GPT 5.6 Sol 20% price reduction”

37. I like the fast releases. 6.1 coming so fast after 6.0 means they found some improvement solid enough for a new rollout.

    Recommends · Sep 2026 · in “GPT-6.1 Sol replaces GPT-6 Sol after just 7 days, with near-Astra intelligence”

38. It didn't solve the problem (for weeks) until I (accidentally) told it which example it happened to know was relevant

    Mixed · Jul 2026 · in “GPT-5.6 used a prompt to close a 30-year gap in convex optimization”

39. I found sol close enough in capability and have only been blocked when my request was an obvious offensive cyber work

    Recommends · Sep 2026 · in “Claude Fable 5.1 and Claude Mythos 5.1”

40. to me this kind of cheating is a better demonstration of intelligence than just attempting the tasks algorithmically

    Recommends · Jun 2026 · in “Previewing GPT‑5.6 Sol: a next-generation model”