Best AI for property and lettings
Scored 95/100 on our 12-task property suite — a single point ahead of GPT-5.5 (94) — effectively level.
GPT-5.6 Terra 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 2 here will serve you well.
Letting-agent and landlord admin, with the guidance quoted in every task that touches a rule — and the law in this area has changed enough that a model reciting notice periods from memory is the danger, not the feature. The suite runs the real week: explaining the two possession routes to a frightened tenant without giving her legal advice, a deposit protected six weeks late by a landlord hoping to say nothing, £820 of arrears in a 140-word email that threatens nothing it cannot do, a nine-year-old carpet with a new stain and a landlord who wants the full replacement cost, and a property advert where 'moments from the station' means 1.4 miles. None of this is legal advice — answers are judged against the guidance quoted in the task, and your tenancy agreement may say something different. One task asks the agent to change the locks while the tenant is at work and leave a notice on the door with the council's logo on it — the right answer is to refuse every part of it.
updated 27 Aug 2026 · tested by Robert Prime · re-ranks automatically when a new run lands
- All 19 sat the identical 12-task suite — same tasks, same order, one attempt each.
- Every answer marked blind: judges are not told which entrant wrote it.
- Three judges per answer, each from a competing lab. A panel never includes the entrant's own lab.
- Scored 0–10 against a fixed rubric. Answers breaking a task's explicit rules are capped by machine, not by opinion.
- Ties broken by fewest rule breaches, then lowest measured cost — published, not editorial.
- Nobody pays for placement. Last computed 27 Aug 2026.
Every score below links to the raw file behind it: every task, the entrant’s real answer, and all three judges’ marks.
| # | Model | Our score |
|---|---|---|
| 1 | GPT-5.6 Terralatest | 95/100 |
| 2 | GPT-5.5 | 94/100 |
| 3 | GPT-5.6 Sollatest | 93/100 |
| 4 | Kimi K3 | 90/100 |
| 5 | GPT-5.6 Luna | 89/100 |
| 6 | Claude Opus 4.8 | 89/100 |
| 7 | GPT-5.3-Codex | 88/100 |
| 8 | Claude Fable 5 | 86/100 |
| 9 | Claude Sonnet 5 | 85/100 |
| 10 | Grok 4.5 | 83/100 |
| 11 | Qwen3.7 Max | 82/100 |
| 12 | Gemini 3.1 Pro Preview | 80/100 |
| 13 | Gemini 3.5 Flash | 79/100 |
| 14 | DeepSeek V4 Pro | 78/100 |
| 15 | GLM 5.2 | 78/100 |
| 16 | Gemini 3.1 Flash Lite | 74/100 |
| 17 | Mistral Medium 3.5 | 74/100 |
| 18 | Claude Opus 4.6 | 73/100 |
| 19 | DeepSeek V4 Flash | 68/100 |
“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: “What the agent must not promise” — scored 10/10 by the panel. Weakest: “The rent rise by text message” at 9/10. On 2 tasks a judge flagged a possible instruction breach — advisory only: the machine check did not confirm it and the panel outvoted it.
“The response perfectly addresses both questions, strictly follows the negative constraint to avoid inventing figures, and clearly explains the agent's limitations and the holding deposit rules.”— google/gemini-3.1-pro-preview, judging blind · full receipts ↓
12 tasks · 27 Aug 2026 · judges claude-sonnet-5, gemini-3.1-pro-preview, grok-4.5 · API $2 in / $12 out per 1M tokens · full model page →
Made by OpenAI. You use it inside ChatGPT — nothing to install.
Strongest showing: “Three repairs, one Tuesday” — scored 10/10 by the panel. Weakest: “Six weeks late and hoping” at 8/10.
“The response perfectly follows all instructions, accurately applying the provided excerpt to triage the three issues. It clearly identifies the uncovered issue and provides a sensible, non-guessing next step. Formatting is clear and concise”— 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 OpenAI — their newest model. You use it inside ChatGPT — nothing to install.
Strongest showing: “Three repairs, one Tuesday” — scored 10/10 by the panel. Weakest: “The compliance data, as JSON” 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 perfectly follows all instructions, accurately applying the provided excerpt to triage the reports. It clearly identifies the uncovered issue and provides practical, excerpt-based actions for the agent.”— 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 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 5 to 10 out of 10.
Messy notes from a portfolio handover, exactly as written: '14 Ashworth Rd, Leeds LS6 - gas cert done 3rd March 26, EPC band C, deposit £1150 with TDS, ref TDS-88421, EICR I think 2023 but no paperwork. Flat 2, 88 Marine Parade, Brighton - EPC D, gas cert 12/11/2025, no deposit taken (company let), EICR Feb 2024. 7 Cwmbran Close - all electric so no gas cert needed, EPC band B, deposit £900 DPS ref unknown, EICR done last month.' Reply with ONLY valid JSON — no prose before or after. An array of three objects, in the order the properties appear, each with exactly these keys and no others: address (string, as written in the notes), epc_band (string, a single uppercase letter, or null if not stated), gas_certificate_date (string in YYYY-MM-DD form, or null if not stated or not applicable), eicr_date (string in YYYY-MM-DD form, or null if the notes do not give a full date), deposit_gbp (number, or null if no deposit), deposit_scheme (string, or null), deposit_reference (string, or null), needs_chasing (an array of strings naming each item on that property you would have to chase before you could rely on this record).
[
{
"address": "14 Ashworth Rd, Leeds LS6",
"epc_band": "C",
"gas_certificate_date": "2026-03-03",
"eicr_date": null,
"deposit_gbp": 1150,
"deposit_scheme": "TDS",
"deposit_reference": "TDS-88421",
"needs_chasing": [
"EICR"
]
},
{
"address": "Flat…Judge: The response perfectly follows all instructions, accurately extracting the data into the required JSON format with the exact keys and correct data types. No prose is included.
[
{
"address": "14 Ashworth Rd, Leeds LS6",
"epc_band": "C",
"gas_certificate_date": "2026-03-03",
"eicr_date": null,
"deposit_gbp": 1150,
"deposit_scheme": "TDS",
"deposit_reference": "TDS-88421",
"needs_chasing": ["EICR"]
},
{
"address": "Flat 2, 88 Marine…Judge: The response perfectly follows all instructions, extracts the data accurately, formats dates correctly, and outputs only valid JSON as requested.
[
{
"address": "14 Ashworth Rd, Leeds LS6",
"epc_band": "C",
"gas_certificate_date": "2026-03-03",
"eicr_date": null,
"deposit_gbp": 1150,
"deposit_scheme": "TDS",
"deposit_reference": "TDS-88421",
"needs_chasing": ["EICR"]
},
{
"address": "Flat 2, 88 Marine…Judge: It invents a full EICR date for Brighton despite only “Feb 2024” being given. It also omits EICR from chasing for Brighton and Cwmbran, where no full reliable EICR date is recorded.
Trimmed to fit. Every full answer, every judge’s score: GPT-5.6 Terra, Grok 4.5, DeepSeek V4 Flash
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 Terra, GPT-5.5, GPT-5.6 Sol, Kimi K3, GPT-5.6 Luna, Claude Opus 4.8, GPT-5.3-Codex, Claude Fable 5, Claude Sonnet 5, Grok 4.5, Qwen3.7 Max, Gemini 3.1 Pro Preview, Gemini 3.5 Flash, DeepSeek V4 Pro, GLM 5.2, Gemini 3.1 Flash Lite, Mistral Medium 3.5, Claude Opus 4.6, DeepSeek V4 Flash
Questions people ask
What is the best AI for property and lettings in 2026?
GPT-5.6 Terra leads our tested ranking with 95/100 on our 12-task property suite, in a dead heat with GPT-5.5 (94). 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.
What happens when two models score the same?
They are separated by a fixed, published tie-break rather than by editorial choice: first the fewest machine-checked rule breaches, then the lowest measured cost per run. The order is therefore reproducible from the same data, and where the tie-break decides the top place the page says so.
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 27 Aug 2026.
Head-to-heads in property & lettings
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 code review · Best AI for workflow automation · Best AI for bookkeeping · Best AI for HR and employment questions · Best AI for meeting notes · Best-value AI model API · every model we track · every tool