KS2 GPS SATs papers 2026: analysis from Sophie Bartlett

Sophie Bartlett returns with her incomparable analysis of the KS2 SATs GPS papers for 2026. Expect the fine-tooth-comb analysis we know and love, her verdict on sneaky preposition questions and even some correct spelling predictions (you could say she had a ‘crystal’ ball for this year’s paper!) 

If you’re after a complete guide to SATs, including when results are released, check out the Ultimate guide to KS2 SATs. 

Share this analysis with your team...

Sign up to our mailing list and get a PowerPoint slides version straight to your inbox.

Get the slides

Another year, another 50 questions testing children on absolutely essential grammatical knowledge that they will definitely continue using throughout their adult lives. Year 6 teachers across the country have finally unclenched their jaws, and thousands of children are currently forgetting what a subordinating conjunction is at an alarming speed.

What’s in this GPS analysis blog for 2026?

The embargo may have lifted, but my head is still firmly buried in test papers and spreadsheets. As always, this analysis looks at:

• the content domain coverage across the papers
• recurring patterns and question styles
• the grammar and spelling choices that made me (and therefore, I assume, everyone else) raise an eyebrow
• and, most importantly, what we can actually learn from it all.

Let’s get stuck in.

Teacher response after the GPS papers

There didn’t seem to be many qualms with the questions themselves - the majority of you (out of over 1,000 responses) thought they were the same level of difficulty, or even easier, than last year’s. 

You also said…

• “Less warm up spellings - deliberately harsh determiner question and adverb question”

• “The exclamation question was RIDICULOUS!”
• “More accessible for our SEND pupils with question types e.g. multiple choice, tick boxes”
• “Spellings were all pretty tricky - all chosen to trip them up”
• “Sneaky ‘which one is NOT’ - just ask the question, don’t try and trip them up!”

Most of you reported that your year 6s found it rather agreeable too! (Are we all using the word ‘agreeable’ a bit more since seeing it in Q10 or is that just me? What a great word.)

So let’s dive into the good stuff. 

GPS Paper 1 analysis

Punctuation makes for the largest proportion of questions in Paper 1 followed by Grammatical terms and word classes.

The English grammar, punctuation and spelling test framework details the profile of marks by content area.

GPS test framework marks by content area

GPS content domains assessed for Paper 1

GPS content domain weighting year-by-year for Paper 1

Please note that whilst these tables only show data from 2022, any analysis commentary will take into account figures from 2017 onwards.

Content domains: the biggest and smallest

G5 remains the most heavily weighted content domain, accounting for almost a third of all questions in Paper 1, with G1 close behind (although this is G1’s lowest proportion to date).

Across the seven content domains there are 51 strands in total, ranging from G1.1 (nouns) to G7.4 (subjunctive). The distribution of marks broadly reflects curriculum size, with G1 and G5 also being the largest domains (12 and 15 strands respectively).

Together, G1 and G5 account for over half of the GPS paper. While this proportion has consistently exceeded the marks needed to pass the paper, it is at its lowest level this year. In theory, this means pupils could still achieve EXS by answering questions solely from these two domains (assuming this year’s pass mark remains below 52%).

These are the individual substrands that are most highly represented:

• Punctuation for parenthesis (G5.9): 22 questions since 2017
• Synonyms and antonyms (G6.1): 18 questions since 2017
• Apostrophes (G5.8): 16 questions since 2017
• Subordinating conjunctions and subordinate clauses (G3.4): 16 questions since 2017

Clauses, synonyms/antonyms and punctuation for parenthesis appeared the most in 2026's paper - more than any other substrand. For the first time, pronouns for cohesion (G1.5) weren’t covered at all.

Overall, there’s very little here that should surprise experienced year 6 teachers. The broad distribution continues the long-established trend of heavily prioritising punctuation and word-level grammar knowledge.

What’s perhaps more interesting is how these domains are being tested. Success depends less on spotting isolated grammatical features and more on understanding how words function within increasingly complex sentence structures.

Content domain coverage only tells us so much. The more interesting story lies in the questions themselves. 

Want the 2026 Reading insights?

Check out Sophie's analysis of the 2026 Reading paper.

Read more

GPS Paper 1: the questions analysis

Starting off strong with EIGHT ‘Tick one’ questions. How dreamy! Even without a clue, everyone has a 25% chance of getting it right (although there’s always a few that will just completely miss a question out, or tick two… but for now, let’s pretend they’ve all been well-trained to read the questions).

I liked question 5 - it may have been tempting to tick the third option if you didn’t read the whole sentence. Properly testing knowledge of using a colon to separate main clauses.

Question 10 was the first to make me ponder. I wouldn’t necessarily expect many year 6s to understand the meaning of ‘agreeable’, but the answer here was certainly achievable through a process of elimination.

Past mistakes have taught me well - the test-writers often hide commands within complex sentences or after adverbials! Question 17 was a good example of this. Identifying the command could also - like the above - have been done via process of elimination, but hopefully many children would know to find the imperative verb after the subordinate clause (even as I’m writing this I can hear Michael Rosen’s voice in the back of my head totally berating the idea that children should have to know this. And I quite agree.)

Now this one is sneaky. ‘After’ can function as many different word types, one of which being a conjunction. In question 18, of course, it’s a preposition (as it’s followed by a phrase) - but I’m sure many children were caught out by this.

Another sneaky one for question 29… the more grammatically astute among us (i.e. nerds) will know that a dangling preposition is to a writer what folding a page corner is to a reader: sacrilege! So not only would it be odd to see a preposition at the end of a sentence, but it also - like the above - isn’t a preposition in this case. Children who have maybe simply memorised a list of prepositions, rather than genuinely understanding their function, are likely to trip up here.

Question 32 was one of my favourites - I do love me a word class question! This paper repeatedly tested children’s understanding of the function of words.

How many of you saw your children hyphenate ‘pasta-sauce’ or ‘garlic-bread’ in question 35? You’re not telling me the test writers didn’t put those open compound nouns in on purpose…

Question 37 fits the recurring pattern in this test of identifying words or phrases in the context of varied sentence structures. Let’s face it, when we teach passive and active, or any types of word class, how often do we present them in simple sentences? Here, the children have had to identify two subjects and objects in a complex sentence - again, really testing their understanding and not just their recall.

More word class identification in question 39. My money is on Year 6s up and down the land circling ‘excited’ without a second thought. ‘Cardiff’ is hopefully a given, with the capital letter in the middle of the sentence; also ‘weather’ as part of a nice, neat noun phrase. It’ll only take those true abstract-noun aficionados to identify the final word as a noun!

Question 41 was one of my least favourites. Why combine converting to active voice with the progressive tense? This seems unnecessarily convoluted for the grand total of one whole mark.

Look, I know we all hate the subjunctive. But let’s lower our swords for a second - don’t you think question 44 was at least a nice one to give them? It’s the classic ‘if I were’, not only at the start of the sentence but as the first option too!

I can hear Michael Rosen in my ear again. Yes, you could write every one of these sentences in question 47 with an exclamation mark at the end, but only one begins with a ‘What’, which is technically what defines an exclamation (in the world of Archaic and Unnecessary Grammar Rules, anyway).

My reaction to many SATs questions is often, “Aw, c’mon, that’s a bit mean isn’t it?” I mean at least they gave them one -ly word as the adverb in question 49, but ‘always’? Sigh.

This paper repeatedly punished surface-level grammar revision and rewarded deeper grammatical understanding.

Want the 2026 maths insights?

Check out Sophie's analysis of the 2026 maths papers.

Read more

Grammar, punctuation and spelling 2026 Paper 2 analysis

The spelling tests contain 20 words – these are split fairly evenly between the Year 3/4 spelling rules and the Year 5/6 spelling rules, although there have always been a few more from Lower Key Stage 2. Across all the spelling papers (Paper 2) so far, 62% of the spellings have come from the Year 3/4 curriculum, and only 38% from the Year 5/6 curriculum.

See the analysis of all the SATs spelling tests below. 

Which spellings appeared? (2017-2026)

• An italic word means it appears under more than one content domain
• A blue word is one of the first three of the test – generally considered slightly easier to spell
• A red word is one of the final three of the test – generally considered slightly harder to spell

Scroll down to scroll across - this table's a biggy!

Spelling analysis: my thoughts

S38 (a Year 3/4 spelling rule) is the most frequently assessed content domain, appearing 17 times across the eight spelling papers, followed by S60 (a Year 5/6 spelling rule), with 11 appearances. Overall, 62% of spellings tested so far come from the LKS2 curriculum, compared with 38% from UKS2.

The spellings are generally considered to increase in difficulty across the test; it is therefore notable that the final four spellings are all drawn from LKS2 spelling rules, which reinforces the importance of secure foundational KS2 spelling knowledge.

This is the first time we’ve seen a repeated word in the spelling test: machine was also used in 2022 (albeit in the singular, as opposed to the plural we saw this year). This follows the Year 3/4 rule of ‘words with the /ʃ/ sound spelt ch’ - in fact, machine is the only word they’ve ever tested children for this rule. I do find this a bit odd considering the spelling appendix gives three other examples (chef, chalet and brochure) and the rule itself has only come up twice anyway. What a niche thing to be concerned about, hey? 

Is now an appropriate time to brag about my spelling predictions again? Both antique and crystal appeared this year - I’ve one-upped myself from two years ago! No pressure for next year. Although if my reading genre predictions are anything to go by (where is that darn poem?!), this was more pure luck than judgement… 

KS2 SATs GPS papers: results

The ‘pass mark’ has historically been between 50-51% (with a random 54% in 2018), whilst the threshold for the higher standard (technically, ‘GDS’ only applies to writing, but a scaled score of 110 is commonly used as the equivalent benchmark) has ranged between 76% - 80% (although we haven’t seen 80% since 2018).

Based on previous year’s data, I’d expect the ‘pass’ mark to be almost exactly half again!

GPS 2026: conclusion

Overall, 2026’s GPS paper felt remarkably balanced. It wouldn’t be SATs without the occasional “Oh, come on now, really?” question (the progressive passive question springs immediately to mind), but on the whole, the paper largely rewarded secure grammatical understanding rather than obscure technicality.

The biggest pattern running throughout the paper was contextual understanding. Time and again, children were asked not merely to identify grammatical features in neat textbook sentences, but to recognise how words and structures functioned within more complex language.

That has implications for teaching too. The children most likely to succeed here probably weren’t the ones who had memorised the longest lists of prepositions or conjunctions, but the ones who had repeatedly encountered grammar in authentic contexts: reading it, discussing it, manipulating it and applying it within real sentences.

As for the spelling paper, the continued dominance of Year 3/4 content once again raises interesting questions about curriculum weighting and assessment priorities. And yes, I will absolutely continue pretending my spelling predictions are based on expertise rather than blind luck.

Until next year’s national debate about whether we think it’s important that a 10-year-old knows that an exclamation begins with ‘what’ or ‘how’ (bring on the curriculum review…)