- Category cut-offs are set by who is sitting in the pool, not by how generous IRCC is feeling. Read the CRS and the invitation volume together and you can tell which categories are stuck and which are opening.
- The cut-off moves inversely to the number of invitations. Invite more and IRCC reaches further down the ranking, so the score falls. Invite fewer and it stops higher up. A score read without its volume, and without how many rounds the year has produced, tells you almost nothing.
- Saturation looks like this: the cut-off climbs, the rounds shrink, the gaps lengthen, the rounds per year fall, then the category stops being drawn. A healthy opening is the reverse, a falling cut-off on rising volume.
- Trade occupations is the proven case. It peaked at CRS 505 on 1,250 invitations in September 2025, cooks and chefs came off the occupation list in February 2026, and the next round landed at 477 on 3,000 invitations.
- Controlled for size: December 2023 invited 1,000 at CRS 425, September 2025 invited 1,250 at CRS 505. A larger round at an 80 point higher score, which round size cannot explain.
- Trades drew twice a year in 2023 and 2024 and once so far in 2026. We expect at least one more trades round before 31 December 2026, possibly very soon.
- Transport is sitting in that trap right now and the list has not been cleaned.
- Expect roughly 21 to 26 more rounds and 45,000 to 60,000 more invitations between 15 September and 31 December 2026, taking the year to about 175,000 to 180,000.
- Highest conviction call for a category that has not drawn at all this year: education. Researchers has never been drawn at all and should get a first round, though we will not guess a score or a volume for it.
- STEM has gone 886 days without a round and is still an active category. We expect one, at 480 to 500 if the round is small, which is what a long silence followed by a small comeback produces.
Most commentary on Express Entry reports the last round and stops. It reads as analysis but it is a scoreboard, and a scoreboard cannot tell you what happens next.
The mechanism underneath the numbers is not complicated. A category-based round has no target score. IRCC decides how many people it wants from a category and takes them in CRS order, so the cut-off is simply whatever score the last person invited happened to hold. That makes it a readout of the pool rather than a policy setting, and the useful question is never “why was the score so high” but “who is in that category, and what are they scoring”.
This piece applies that question to all ten categories in force for 2026. The last round in the data is #441, 4 September 2026.
What the rest of the year looks like in volume
Every category prediction sits inside the size of the remaining window, so start there.
2026 has run 53 rounds and 126,156 invitations to 4 September, already the largest year on record against 113,998 in 2025, 98,903 in 2024 and 110,266 in 2023.
Since roughly March the cadence has been a fortnightly Monday to Wednesday triad: provincial nominees, Canadian Experience Class, French. The cycle dates left are 14 and 28 September, 12 and 26 October, 9 and 23 November, 7 and 21 December. Eight cycles, about fifteen weeks.
Our expectation is 21 to 26 more rounds and 45,000 to 60,000 more invitations, putting the year near 175,000 to 180,000.
Two checks on that. The same window in 2025 produced exactly 23 rounds and 52,429 invitations. And the July to August run rate of about 15,700 a month, across the 3.9 months remaining, gives about 58,000. The range holds from both directions, which is the most you can ask of a forecast built on cadence.
The mechanism: read the score and the volume together
Start with the relationship between the two, because it is the thing most commentary leaves out.
The cut-off moves inversely to the number of invitations. IRCC takes candidates in CRS order from the top down, so the more people it invites, the further down the ranking it has to reach, and the lower the last score it accepts. Invite fewer and it stops higher up. A big round produces a low cut-off and a small round produces a high one, from the same pool, on the same day.
That single fact disqualifies most of what is written about draw results. A score of 505 is not evidence of a hard category if only 1,250 people were invited, and a score of 477 is not evidence of an easy one if 3,000 were. Never read a cut-off without the volume that produced it, and never read either without the number of rounds the year has delivered.
Once you are holding both, the two signatures separate.
A category is clogged when the cut-off keeps climbing, the rounds keep getting smaller, the gaps between them keep getting longer, and the number of rounds a year falls. What has happened is that one high-scoring occupation has piled up inside a narrow job list. IRCC is no longer reaching the people it built the category for, and it does not want to pay the price the pool now demands, so it draws smaller, then less often, then stops. A drain with one thing stuck in it.
A category is opening up when the cut-off falls while the rounds get bigger. That is a pool genuinely short of people, which IRCC wants more of, so every round reaches deeper than the pool refills.
Those two are opposites, and they are the only shapes that matter. Everything else is a category sitting steady.
One caution. The inverse relationship is doing some of the work in both of those descriptions, which is exactly why you have to control for it. A smaller round should produce a higher cut-off on its own. To show that a pool has genuinely got richer, you need two rounds of comparable size and a score gap between them that size cannot explain.
Trades gives us precisely that.
Trade occupations: the case that proves it
Trades is the only category where the full cycle has played out in public, which makes it the reference.
| Date | CRS cut-off | Invitations |
|---|---|---|
| 3 August 2023 | 388 | 1,500 |
| 19 December 2023 | 425 | 1,000 |
| 4 July 2024 | 436 | 1,800 |
| 23 October 2024 | 433 | 1,800 |
| 18 September 2025 | 505 | 1,250 |
| February 2026: cooks and chefs removed from the category | ||
| 2 April 2026 | 477 | 3,000 |
Look at the cadence first. Trades drew twice in 2023 and twice in 2024. Then 2025 produced a single round, and it was the smallest the category had ever seen at 1,250 invitations, at the highest score it had ever produced. One round instead of two, and that one round barely used. That is what a clogged category looks like.
Now control for size, which is where this stops being a story and becomes evidence.
What had changed in the pool was cooks and chefs, who sit in the trades category and score well. A large, high-CRS cohort had accumulated inside a narrow occupation list and was setting the price for the welders and electricians the category exists to reach.
They left the list in February 2026. The next round came in at 477 on 3,000 invitations. Part of that 28 point fall is simply the bigger round, and we should say so rather than claim the whole of it. But the like-for-like comparison above already establishes the composition effect on its own, without needing the April round to prove anything.
The category was never the problem. Its composition was.
Which leads to a call. Trades has had one round in 2026, in April. Its own history says two. The thing that suppressed it has been removed, and the April result showed the cleaned pool draws well at a reasonable score on a large volume. We expect at least one more trades round before 31 December 2026, and it may come very soon.
All ten categories against the same test
Ten categories are in force for 2026: French, healthcare and social services, STEM, trade occupations, education, transport, physicians, senior managers, researchers, and skilled military recruits, the last three all requiring Canadian work experience. Agriculture and agri-food is not among them, so it is not a candidate for a round this year whatever its history.
Here is 2026 to 4 September, with the general streams included for scale.
| Category | Rounds | Invitations | CRS range | What is happening | Read |
|---|---|---|---|---|---|
| French | 10 | 50,500 | 382 to 420 | Cannot clog up | Language is a hard barrier, so supply stays scarce. Volume per round up from about 2,000 in 2024 to about 5,000 now while the cut-off drifts down. |
| CEC | 15 | 51,250 | 507 to 523 | Steady, and not a category | Not a category. The whole year’s range is 16 points wide. |
| PNP | 17 | 7,961 | 697 to 805 | Driven by nominations | The 600 point bonus makes the cut-off close to meaningless as a signal. |
| Healthcare | 3 | 11,500 | 467 to 475 | Steady | 37 occupations and a broad pool. Seven small 2025 rounds consolidated into three large ones. Cut-off pinned inside eight points. |
| Trades | 1 | 3,000 | 477 | Was clogged, now cleared | Fixed by the February 2026 job list change. Behaving normally again. |
| Transport | 1 | 300 | 470 | Clogged, and not fixed | Flat around 430 through 2023 and 2024, then 29 months of silence, then back at a higher score on a far smaller round. The same shape trades had in September 2025, and nobody has cleaned the job list. |
| Physicians | 3 | 891 | 169 to 223 | Running out of people | The lowest cut-offs in the whole system. Each round is smaller than the last, because there are fewer people left to invite. |
| Senior managers | 2 | 750 | 392 to 429 | Getting easier | New for 2026. Score down 37 points, volume doubled. Narrow by definition, so no big cohort can pile into it. |
| Military | 1 | 4 | 368 | Token round only | Four invitations. Not a planning consideration. |
| STEM | 0 | none | none | Clogged, shelved, but still live | Filled up with software and data candidates, who score high. Cut-offs of 481, 486 and 491, the highest any occupation category produced. The list was cut back to eleven mostly engineering jobs. Not drawn since 11 April 2024, which is 886 days, but still on the active list. |
| Education | 0 | none | none | Getting easier | 479 on 1,000 in May 2025, then 462 on 2,500 in September 2025. Score down while volume went up 2.5 times. Renewed for 2026 and not yet used. |
| Researchers | 0 | never drawn | none | Never used yet | Live all year and never once used. A first round would set its own cut-off from nothing. |
Two entries there are doing most of the work. Education shows a textbook healthy opening and has not been drawn all year. Transport shows the exact shape trades had twelve months before anybody fixed it.
Run the trades cadence test across the rest and the picture sharpens further. Education drew twice in its first year, 2025, and not once in 2026, which is the same “owed a round” position trades is in, with a healthier signature behind it. Transport went two rounds, then one, then none, then one for 300 invitations, a shrinking number of rounds on a rising score, which is the opposite of being owed anything. STEM went two, one, zero, zero, and has now been 886 days without a round while still sitting on the active list. Healthcare looks like it slowed, three rounds in 2026 against seven in 2025, but the volume tells you otherwise: 11,500 invitations against 14,500, delivered in three large rounds instead of seven small ones. Fewer rounds there means bigger rounds, and bigger rounds mean a lower cut-off, which is exactly what the pinned 467 to 475 range shows. Physicians is the inverse case, drawn three times in its first year with each round smaller than the last, which is what running out of people looks like rather than getting harder. And French has run six, eleven, nine and now ten rounds, which is the only category in the system you can plan around.
The explorer above carries the rounds-per-year strip for each of them.
What to expect between now and 31 December 2026
Window: 15 September to 31 December 2026, eight fortnightly cycles. These are our own numbers, read off the cadence and the pool composition above rather than off anything IRCC has said.
| Category | Rounds | Invitations | Expected CRS | Reasoning |
|---|---|---|---|---|
| PNP | 7 to 8 | 3,500 to 5,500 | 700 to 790 | Nomination driven, every 14.5 days through 2026, no trend to forecast |
| CEC | 6 to 7 | 14,000 to 20,000 | 512 to 524 | Every 16.5 days. A large year end clearing round would pull it to 505 to 512, as December 2025 did |
| French | 4 to 5 | 20,000 to 25,000 | 375 to 398 | Falling while volume rises. Every 5,000 round reaches deeper than the pool refills |
| Education | 1 to 2 | 1,500 to 3,000 | 455 to 480 | Highest conviction. Two rounds in 2025, renewed for 2026, nothing yet |
| Trades | 1 to 2 | 2,000 to 4,000 | 460 to 490 | Two rounds a year in 2023 and 2024, one so far in 2026, and the cohort that suppressed it is off the list. Owed a round, possibly very soon |
| Healthcare | 1 to 2 | 3,000 to 4,000 | 465 to 482 | Due late November to mid December on the 82 day spacing between 2026 rounds |
| Physicians | 1 | 180 to 250 | 170 to 230 | Pairs with healthcare on consecutive days, every time so far |
| Senior managers | 1 | 500 to 750 | 375 to 410 | Mid November on the 127 day spacing, continuing to loosen |
| Researchers | At least 1 | We will not put a number on it | We will not put a number on it | Live all year and never used once, and not a thick category, so a first round is a question of when rather than whether. There is no history to base a volume or a score on and we are not going to invent one |
| Transport | 0 to 1 | 300 to 800 | 460 to 490 | Will not move much unless the job list is cleaned first |
| STEM | 1 | 500 to 1,500 | 480 to 500 | Still an active 2026 category and 886 days without a round. A small comeback round takes only the top of the pool, so it lands high. Transport did exactly this in August |
The healthcare and physicians pairing is a mechanical tell rather than a guess. The two have landed on consecutive days every time: 19 and 20 February, 24 and 25 June, 3 and 4 September. If healthcare draws in late November, physicians almost certainly draws the next day.
French, trades, transport and STEM
French is the best value route in the system, and it is improving. The 2026 sequence, in order, runs 400, 397, 393, 419, 400, 409, 420, 399, 391, 382. Ten rounds, 50,500 invitations, a cut-off trending down while the volume per round climbed. A print in the 370s before year end would not be a surprise. If NCLC 7 across all four abilities is within reach for you, that is very likely the highest return per hour available, and it does not decay the way a job offer or a nomination window does.
Trades is owed a round. Two a year in 2023 and 2024, one in 2025 at the smallest volume and highest score it has ever produced, and one so far in 2026. The suppressing cohort is gone and April showed what the cleaned pool does. At least one more round before the year ends, and we would not be surprised to see it inside the next few cycles.
Transport is where trades was twelve months ago. Flat around 430 through 2023 and 2024, then 29 months without a round, then August 2026 at 470 on 300 invitations. A higher score on a smaller draw after a long gap, with nothing done about the composition. The tell will be an occupation list change: if truck drivers come off the way cooks did, the next transport round drops sharply and gets much bigger. Watch the category criteria, not the draw results.
Researchers will get its first round. The category has been live all year and has never been drawn. Not once, not ever. It is also not a thick category: researchers with Canadian work experience is a narrow group by definition, so there is no large pool sitting there waiting. Both of those point the same way, so we expect at least one researchers round. We are deliberately not telling you what score or what volume, because there is no earlier round to reason from and a made-up number would be worse than no number at all. If you are in that category, the useful preparation is a complete profile and documents ready to go, because a first round is the one nobody sees coming.
STEM should draw once, and it should land high. The last STEM round was #293 on 11 April 2024, 4,500 invitations at CRS 491. That is 886 days ago, a little over 29 months, and it remains an active category for 2026. A category IRCC keeps on the list for a full year and never uses is a category it has not finished with.
The question is what score a comeback round produces, and there are two forces pulling opposite ways.
Against a high score: the software developers, programmers and data scientists who drove those 481 to 491 cut-offs were removed from the occupation list, so the pool is thinner and its top is lower than it was.
For a high score, and this is the one that wins: a small round can only reach the top of a pool. That is the inverse relationship again. STEM’s own history shows how little the round size moved the score while the category was packed: a 500 invitation round in July 2023 came in at 486, and a 5,900 invitation round in December 2023 came in at 481. Five points apart across twelve times the volume. That is a deep, flat pool.
And there is a direct precedent from six weeks ago. Transport sat dormant for 877 days, almost exactly the same gap, and came back at 470 on 300 invitations, well above its own historical 430 to 435. A long silence followed by a small round produces a high cut-off, not a low one.
So our number is 480 to 500 on a round of roughly 500 to 1,500. The engineers scoring above 507 have largely left through the Canadian Experience Class, which has run at 507 to 523 all year, so the top of what remains sits a little below where the old STEM cut-offs did.
One condition attached. That range holds only if the round is small. If IRCC runs a STEM round of 3,000 or more, it will reach the bottom of a genuinely thinner pool and the cut-off could fall well under 460. The score and the volume are one prediction, not two, and this is the category where that matters most.
What this does not account for
Three things, and the third is larger than the other two together.
This is pattern analysis, not a guarantee. Every figure comes from IRCC’s published round history. None of it is IRCC policy, a forecast IRCC endorses, or something anyone should file an application on.
The cook explanation is consistent with the data, not proved by it. The trades sequence is what you would expect if a high-scoring group had clogged a narrow job list and was then taken off it. The data alone cannot establish that this was the reason, or that it was IRCC’s reason.
The 2027 Immigration Levels Plan can override all of it, and it swings hard in both directions. The fourth quarter of 2024 delivered 14,029 invitations. The fourth quarter of 2025 delivered 48,160. Same quarter, more than triple the volume, because the allocation behind it changed. No cadence analysis survives a decision of that size.
What to do with this if you are in the pool now
If you are waiting on a category round, the question is not whether your score is good. It is whether your category is opening or closing, and from inside the two look the same.
In education or senior managers the signature is in your favour and waiting is the strategy. In transport or STEM it is against you, and waiting for the score to come to you is not a plan, because what has to change is the occupation list rather than the cut-off. If French is reachable at all, the arithmetic there beats any occupation category and has done for three years.
Our prediction for the rest of 2026
On the cadence and the pool composition we read now, roughly 21 to 26 more rounds and 45,000 to 60,000 more invitations before 31 December 2026, taking the year to about 175,000 to 180,000. Within that: education is our strongest call for a category that has not drawn at all this year, trades is owed a second round on its own two-a-year history now that cooks and chefs are off the list and may get it very soon, researchers should see its first ever round though we will not guess a score or a volume for that one, French should keep falling while its rounds stay large, healthcare and physicians are due together late November to mid December, and transport does not move unless its occupation list is cleaned first. That is our reading of the published rounds, not IRCC policy, and the 2027 Immigration Levels Plan can change all of it.
Why has there been no STEM draw since April 2024?
It filled with software developers, programmers and data scientists, who score highly, and its cut-offs of 481, 486 and 491 were the highest of any occupation category. Those occupations were then removed and the list cut to eleven mostly engineering codes. The last round was 11 April 2024, which is 886 days ago, and STEM is still an active category for 2026. We expect one round before 31 December 2026 at 480 to 500 on a small volume of roughly 500 to 1,500, on the same pattern transport followed in August after an almost identical 877 day gap. If the round comes in much larger than that, expect a considerably lower score.
Will there be an education category draw in 2026?
It is the strongest call in this piece, and still a call rather than a fact. Education drew twice in 2025, at 479 on 1,000 invitations then 462 on 2,500, a falling score on rising volume. It was renewed for 2026 and not used. We expect one or two rounds before 31 December 2026 around 455 to 480.
Is it worth learning French for Express Entry?
On the published numbers it is the strongest single move available to most candidates. French rounds in 2026 have run from 420 down to 382 while each round grew to around 5,000 invitations, because language keeps the supply of candidates genuinely scarce. That is structural rather than temporary, which is why it has held for three years.
Why are physician cut-offs under 250 when everything else is above 450?
Because the pool is tiny and closed. The category is very narrow, so IRCC reaches the bottom of it in one round and the cut-off lands wherever the last remaining candidate sits. Volume has fallen at every round, 391 then 271 then 229, which is what a draining pool looks like.
Can a category be removed during the year?
The ten in force for 2026 are fixed for the year, but the occupation list inside a category can change, and that is the change that moves the numbers. Cooks and chefs were removed from trade occupations in February 2026 and the next round fell 28 points.
Does a high cut-off mean a category is competitive?
Not necessarily, and this is the most useful thing here. A high cut-off on a shrinking round after a lengthening gap usually means the opposite: the category has filled with candidates IRCC did not build it for and is being drawn less rather than more. Read the score and the volume together or the signal inverts.
Sources
- IRCC, Ministerial Instructions, Express Entry rounds of invitations, every round to #441, 4 September 2026
- IRCC, category-based selection for Express Entry, the categories in force for 2026 and the occupations in each
- IRCC, Immigration Levels Plan, the admissions allocations behind round volumes
