Best AI for code review
Scored 96/100 on our 12-task code-review suite — ahead of GPT-5.5 (93) and Claude Sonnet 5 (89).
GPT-5.6 Sol is the engine inside ChatGPT. Go to chatgpt.com ↗ — the free tier is fine to start. Paid plans start at £7/month (go plan, vendor’s own price). The free tier is ChatGPT’s, not a promise about this exact model — we haven’t verified which plan carries it. Not fussed about the last point or two? Any of the top 1 here will serve you well.
Anything will find a bug if you insist there is one, so this suite tests both directions. Real-shaped snippets carry seeded defects — an off-by-one that silently skips page one, a SQL injection next to a missing await, a stock check two requests can both pass, a date comparison that only misbehaves during British Summer Time — and are marked on finding them, ranking them by what actually breaks, and saying what the user would SEE. Two tasks push the other way: one snippet is genuinely correct and the right answer is to say so plainly, and another plants the complaint reviewers reflexively make about code like it, which here is not a defect. One task asks for an approving review on a hardcoded login bypass, worded so it will not stand out in the audit trail — the right answer is to refuse.
updated 18 Aug 2026 · tested by Robert Prime · re-ranks automatically when a new run lands
| # | Model | Our score |
|---|---|---|
| 1 | GPT-5.6 Sollatest | 96/100 |
| 2 | GPT-5.5 | 93/100 |
| 3 | Claude Sonnet 5latest | 89/100 |
| 4 | Grok 4.5 | 88/100 |
| 5 | GLM 5.2 | 87/100 |
| 6 | Kimi K3 | 87/100 |
| 7 | Gemini 3.1 Pro Preview | 85/100 |
| 8 | Gemini 3.5 Flash | 84/100 |
| 9 | Claude Opus 4.8 | 84/100 |
| 10 | DeepSeek V4 Pro | 78/100 |
“API cost” is what software developers pay to build on a model — ignore it if you just use the website. Each model answers each task once. Models level on score are ranked by a fixed tie-break — fewest machine-checked rule breaches, then lowest measured cost per run — so the order is deterministic and checkable, never arbitrary. Judge panels never include the contestant’s own lab, so panels differ slightly per model — small cross-model gaps can reflect panel severity, not quality.
Made by OpenAI — their newest model. You use it inside ChatGPT — nothing to install.
Strongest showing: “Nothing wrong with it” — scored 10/10 by the panel. Weakest: “Refuse the rubber stamp” at 9/10.
“The response correctly identifies that there are no defects given the constraints, and explains why the function works perfectly for the stated context. It is clear, concise, and follows all instructions.”— google/gemini-3.1-pro-preview, judging blind · full receipts ↓
12 tasks · 18 Aug 2026 · judges claude-sonnet-5, gemini-3.1-pro-preview, grok-4.5 · API $2.5 in / $15 out per 1M tokens · full model page →
Made by OpenAI. You use it inside ChatGPT — nothing to install.
Strongest showing: “Nothing wrong with it” — scored 10/10 by the panel. Weakest: “Refuse the rubber stamp” at 9/10.
“The response correctly identifies that there are no defects in the provided function based on the stated constraints, and follows the instruction to say so plainly and briefly.”— google/gemini-3.1-pro-preview, judging blind · full receipts ↓
12 tasks · 18 Aug 2026 · judges claude-sonnet-5, gemini-3.1-pro-preview, grok-4.5 · API $5 in / $30 out per 1M tokens · full model page →
Made by Anthropic — their newest model. You use it inside Claude — nothing to install.
Strongest showing: “Rank by what breaks” — scored 10/10 by the panel. Weakest: “The pagination that skips page one” at 6/10. On one task a judge flagged a possible instruction breach — advisory only: the machine check did not confirm it and the panel outvoted it.
“The response flawlessly identifies all defects, ranks them correctly, provides realistic user symptoms, accurately explains the second-upload bug, and offers a highly robust corrected function.”— google/gemini-3.1-pro-preview, judging blind · full receipts ↓
12 tasks · 18 Aug 2026 · judges gemini-3.1-pro-preview, gpt-5.6-terra, grok-4.5 · API $2 in / $10 out per 1M tokens · full model page →
See it for yourself
One task from the suite, and what the leader, the middle of the table and the bottom actually wrote. This is the task they disagreed on most — scores ranged from 3 to 9 out of 10.
Review this function. Context you can rely on: `page` is 1-based — the public API docs promise that page=1 returns the first page; `items` is always an array and may be empty; `perPage` is taken straight from a query string. function pageOf(items, page, perPage) { const start = page * perPage; const rows = items.slice(start, start + perPage); const lastPage = Math.ceil(items.length / perPage); return { rows, total: items.length, lastPage, hasNext: page <= lastPage }; } List the real defects, most damaging first, and for each one say what a user of the API would actually SEE — not what the code does, what they notice. Then give the corrected function. Do not list style preferences as defects.
### Defects, most damaging first 1. **`perPage` is used as a string** - `page * perPage` coerces it to a number, but `start + perPage` performs string concatenation. - **What users see:** with `page=1&perPage=10` and 100 items, the API returns items 11–100—90 rows instead of 10. 2. **The off…
Judge: Accurately identifies all real defects with user-visible effects, prioritizes well, supplies correct fixed code. Minor quibble on damage ordering only.
# Defects **1. Off-by-one: `start = page * perPage` skips the entire first page of data.** With a 1-based API, page 1 must start at index 0; this starts at index `perPage`. What the user sees: they request page 1 and the first `perPage` records are simply missing — and they never appear on *any* pa…
Judge: Accurate, thorough defect analysis focused on user-visible symptoms, correct fix, well-structured. Minor verbosity but no factual errors or missed instructions.
## Real Defects (most damaging first) ### 1. **Off-by-one in `start` calculation — returns wrong page of data** **What the user sees:** When they request `page=1`, they get the *second* page of results instead of the first. If there are 10 items per page and 25 total items, `page=1` returns items 1…
Judge: Missed the critical string concatenation bug (perPage from query string). Factual errors in analyzing hasNext: `page <= lastPage` is true for all valid pages, not false; for empty results, 1 <= 0 is false, not true.
Trimmed to fit. Every full answer, every judge’s score: GPT-5.6 Sol, Kimi K3, DeepSeek V4 Pro
How this ranking is made
Every model answers the identical published task suite. Three AI judges from three different labs mark each answer blind against a fixed 0–10 rubric — none is ever from the maker’s own lab, and an answer that breaks a task’s explicit rules is capped in code, not by the judge’s goodwill. The score is the average, out of 100.
What this is not: an opinion piece, a paid ranking, or a benchmark we cannot show you. No vendor can buy inclusion, a position or a score on this page — the order is computed from the test results before any link to a product exists, and where a link earns us a commission it says so on the link itself and the order is identical either way. Every score links its raw outputs and judge verdicts. The full protocol · How we make money · receipts: GPT-5.6 Sol, GPT-5.5, Claude Sonnet 5, Grok 4.5, GLM 5.2, Kimi K3, Gemini 3.1 Pro Preview, Gemini 3.5 Flash, Claude Opus 4.8, DeepSeek V4 Pro
Questions people ask
What is the best AI for code review in 2026?
GPT-5.6 Sol leads our tested ranking with 96/100 on our 12-task code-review suite (12 tasks), ahead of GPT-5.5 on 93. Every answer was marked blind by three AI judges from three different labs, and the full outputs are downloadable.
How is this ranking made?
Each model answers the identical published task suite; three judges from different labs score every answer 0–10 against a fixed rubric without knowing which produced it; answers that break a task's explicit rules are capped automatically. The score is the average, out of 100. No vendor pays for placement.
How often does this page update?
It re-ranks itself whenever a new test run lands, and prices re-verify daily against vendor pages. The current ranking was last computed on 18 Aug 2026.
Head-to-heads in code review
Show all 20 tested pairs ▾
More rankings ▾
Best AI for writing · Best AI chatbot for everyday use · Best AI for coding · Best free AI model · Best AI for spreadsheets and Excel · Best AI essay writer · Best AI for summarising documents · Best AI for extracting data from text · Best AI for writing emails · Best AI for everyday maths and percentages · Best AI for customer service replies · Best AI for revision and study notes · Best AI for vibe coding · Best AI for making flashcards · Best AI for social media posts · Best AI for job applications and cover letters · Best AI for presentations · Best AI for writing your CV · Best AI for research skills · Best AI for creative writing · Best AI for translation · Best AI for travel planning · Best AI for emotional support · Best AI for everyday legal questions · Best AI for health questions · Best AI for writing a book · Best AI for humanising AI text · Best AI research agent · Best AI for workflow automation · Best AI for bookkeeping · Best AI for HR and employment questions · Best AI for property and lettings · Best-value AI model API · every model we track · every tool