Best AI for coding / head-to-head
GPT-5.5 vs Kimi K3for coding
GPT-5.5 wins, 91–89
Same 18 coding tasks, marked blind by three rival labs. GPT-5.5 took 7 tasks, Kimi K3 took 4, 7 tied. Tested 16 Aug 2026.
Where they differed most
Write a Python function parse_duration(s) that converts strings like '1h30m', '45s', '2h', '90m', '1h2m3s' into total seconds. Raise ValueError on anything malformed. Include three assert-based tests, one of which covers a malformed input.
```python
import re
def parse_duration(s):
"""
Convert a duration string like '1h30m', '45s', '2h', '90m',
or '1h2m3s' into total seconds.
Raises ValueError if the input is malformed.
"""
if not isinstance(s, str):
raise ValueError("duration must be a string")
match = re.fullmatch(r"(?:(\d+)h)?(?:(\d+)m)?(?:(\d+)s)?", s)
if not match or not any(match.groups()):
raise ValueError(f"malformed duration: {s!r}")
hours, minutes, seconds = match.groups()
return (
int(hours or 0) * 3…```python
import re
_FACTORS = {'h': 3600, 'm': 60, 's': 1}
def parse_duration(s):
"""Convert a duration like '1h30m' or '45s' into total seconds.
Raises ValueError if the string isn't one or more <digits><unit> pairs.
"""
if not isinstance(s, str) or not re.fullmatch(r'(?:\d+[hms])+', s):
raise ValueError(f"malformed duration: {s!r}")
return sum(int(num) * _FACTORS[unit]
for num, unit in re.findall(r'(\d+)([hms])', s))
# Tests
assert parse_duration('1h2m3s') == 3723 # 3600 + 120 + 3
assert pars…Task by task
| Task | GPT-5.5 | Kimi K3 |
|---|---|---|
| Duration parser | 8 | 10 |
| Find the bug | 10 | 10 |
| SQL without window functions | 10 | 10 |
| Infinite useEffect | 10 | 9 |
| Typed debounce | 9 | 10 |
| Leftmost binary search | 10 | 8 |
| Security review | 10 | 9 |
| Safe migration | 9 | 9 |
| Fix the code not the test | 9 | 9 |
| Race condition | 10 | 9 |
| Retry with backoff | 5 | 5 |
| Recursive type | 10 | 8 |
| Bash one-liner | 9 | 9 |
| Explain and cost | 10 | 9 |
| Regex with limits | 8 | 10 |
| Callback to async | 9 | 8 |
| Diagnose from a trace | 8 | 9 |
| Idempotency | 9 | 9 |
Full receipts: GPT-5.5, Kimi K3 · judges claude-sonnet-5, gemini-3.1-pro-preview, grok-4.5
Questions people ask
Which is better for coding: GPT-5.5 or Kimi K3?
GPT-5.5 — it scored 91/100 against 89/100 on our 18-task coding suite, winning 7 tasks to 4 with 7 tied. Every answer was marked blind by three judges from three rival AI labs.
How was this tested?
Both models answered the identical published coding tasks. Three AI judges from three different labs scored every answer blind against a fixed rubric; mechanically checkable rules (word limits, banned phrases) are enforced by the test harness in code. The raw outputs and judge verdicts are downloadable.
More coding head-to-heads: GPT-5.5 vs GPT-5.6 Luna · GPT-5.6 Luna vs Kimi K3 · GPT-5.5 vs GPT-5.6 Sol · GPT-5.6 Sol vs Kimi K3 · GPT-5.5 vs GPT-5.6 Terra · GPT-5.6 Terra vs Kimi K3
Full ranking: Best AI for coding · model pages: GPT-5.5, Kimi K3