Kimi K2.6: what 38 developers actually say.
38 Hacker News opinions from 31 developers, Nov 2025 – Nov 2026. 31 come from people who say they use it. Here is what they praise, what they complain about, and who switched.
The case for Kimi K2.6, in one paragraph.
Kimi K2.6 earns 70% approval across 38 opinions. Developers are divided on it.
Pick it if you
- Need solid code on tightly scoped tasks
- Already have a detailed plan from a Claude model
- Do security or binary analysis work
Look elsewhere if you
- Write Zig
- Run long agentic workflows
- Need fast local inference
Where it wins and loses.
Mentions across 38 opinions. Bars are counts, not votes; a tool can be both praised and criticized on the same topic.
View data table
| Topic | Praise | Complaints | Mentions |
|---|---|---|---|
| Code quality | 11 | 3 | 14 |
How opinion has moved.
Monthly approval, Apr 2026 – Jul 2026. Bars show opinion volume; low-volume months are still included but faded.
View data table
| Month | Approval | Opinions |
|---|---|---|
| Apr 2026 | 90% | 10 |
| May 2026 | 58% | 13 |
| Jun 2026 | 64% | 11 |
| Jul 2026 | 75% | 4 |
What developers say.
Summaries are paraphrased. Each links to the original Hacker News comment.
Praise
4 selectedKimi K2.6 is very competent at mapping binary flows and patterns, though it is slow.
Uses Kimi K2.6 for much work and finds it extremely effective.
Kimi K2.6 finds vulnerabilities better than Opus models, despite some false positives.
Had Kimi K2.6 build a Rust diff tool in one shot, which they still use.
Complaints
4 selectedKimi K2.6 struggles to get reference counting right in Zig, even on the happy path.
Kimi K2.6 gave no good results on their local agentic tasks.
Kimi K2.6 keeps writing C-style Zig, ignoring defer and opaque enums even when told explicitly.
Ran Kimi K2.6 locally at Q4 for SIMD work, but found it too slow to be practical.
Questions developers ask.
Is Kimi K2.6 worth it?
Is Kimi K2.6 better than Claude Opus?
Is Kimi K2.6 good for agentic coding?
The data behind this page.
Kimi K2.6 scores 70% approval across 38 opinions from 31 authors, and 31 of those opinions are firsthand. Code quality is the strongest signal at 11 positive to 3 negative, and cost is 8 positive to 1 negative. Approval fell from 90 in April to 58 in May, then moved to 64 in June and 75 in July, though July has only 4 opinions. Speed and reliability draw more criticism than praise, with reliability at 1 positive to 2 negative.
Every number comes from public Hacker News comments collected through the HN Search API. A language model tagged each comment for sentiment, first-hand use, topic and switches; a human spot-checks samples. No vendor can pay for placement.
Get the monthly report
Which AI coding tools gained or lost developer approval on Hacker News, and why people switched. One email a month.