Research notes

TEF Canada: Listening (CO) and Reading (CE), item by item

Research date: 4 November 2026. Target: NCLC 7 in CO and CE. That means 434 or more in the "Score /699" column of the attestation. IRCC reads the "Équivalence ancien score" column instead, where NCLC 7 starts at 249/360 for listening and 207/300 for reading. Both columns appear on the same attestation and lead to the same NCLC level (NCLC table, IRCC, sample attestation).

0. Bottom line

1. Half of the listening test is short everyday audio with closed question types. Questions 1 to 20 are four kinds of task. Match a dialogue to a drawing. Say why an announcement is broadcast. Say why a caller left a voicemail. Say whether a speaker is for, for under conditions, or against. These 20 questions are the cheapest listening points on the test. Bank them first (CO sample).
2. The official tutorial dated 26-08-25 shows two changes that make listening easier than older prep material suggests. First, vox-pop (micro-trottoir) questions now have 3 options instead of 4: the "no opinion" option is gone. Second, radio interviews are played twice. All other audio is still played once (tutorial title/date, vox-pop screen, interview screen). Older practice sets with 4 vox-pop options and single-play interviews are out of date on both points.
3. Reading has 18 low-level questions, and you can move freely within its 60 minutes. The 18 are Q1-13 and Q18-22: everyday documents, gap sentences, scanning short texts and one graph. Do them first and fast. Spend the saved time on the 10 administrative/professional questions and the 8 press questions (CE sample, livret 2026 p.14).
4. The conversion from correct answers to score is not published. Plan for 28/40 (70%) correct in each test on official-format practice. Somewhere between 25 and 28 is probably enough for 434, but no official figure exists (section 3).
5. Learn first: vox-pop stance markers; purpose verbs and voicemail/announcement scripts; everyday-scene words plus numbers, times and prices; names of document types; graph-trend words; collocations for the gap sentences; formal email phrases; reporting and attitude verbs; connectors. Grammar: infinitive of purpose, negation (including ne…que, personne, aucun), telling done from promised (past vs future), conditional for hedged claims, relative pronouns with en and y, and the passive.

1. Sources and how much to trust them

• Official sources (format and scoring), all from Le français des affaires (CCI Paris Île-de-France):
  • sample booklets for CO and CE, uploaded October 2024 (CO sample, CE sample)
  • the interactive CO tutorial, whose module title is dated 26-08-25 (CO tutorial), and the CE tutorial (CE tutorial data)
  • the CO preparation page of 11 Dec 2023 (CO prep page) and the December 2023 change notice (2023 changes)
  • the TEF Canada presentation and results pages (presentation, results FR)
  • the TEF to NCLC table (NCLC table) and the sample attestation (sample attestation)
  • the administrator booklets of October 2025 and April 2026 (livret 2025, livret 2026).
  • IRCC's Express Entry language-testing page (IRCC).
• Prep books:
  • The official preparation page lists the Hachette FLE book TEF nouvelle édition – livre d'entraînement (preparation page). It was published 25/04/2023, and the publisher page gives only an outline: 330+ activities, one mock test per paper, transcripts and answer keys (Hachette book).
  • CLE International's ABC TEF was published 6 May 2025 (ABC TEF listing; seen only in a search result, the page itself did not load).
  • I could not open the table of contents of either book, so no counts below come from books.
• Candidate reports: I found no candidate-recalled CO or CE question bank for the current TEF Canada (format of 11 Dec 2023 onward).
  • Recall banks exist for TCF writing and speaking. Nothing comparable surfaced for TEF listening or reading.
  • The one forum thread found is from 2015-2018, before the current format (pvtistes thread).
  • Reddit could not be fetched from here.
  • The web-search budget for this session ran out before YouTube and Facebook compilations could be searched further.
  • So "what repeats" below is measured in two ways:
    • (a) the official question counts per task type, which every test form repeats by design;
    • (b) the 40 official sample items (19 CO + 21 CE). The 2025 tutorials reuse the same items.
  • Third-party prep guides are used only for tactics and traps, and are labelled as third-party.
• The official booklets say the sample items are not a full test and do not reflect the real difficulty (CO sample p.2, CE sample p.2).

2. Format facts that change how you study

What changed in December 2023 (2023 changes, high confidence):

• Reading:
  • Removed: text reordering and synonym sentences (old sections C and D).
  • Modified: gapped texts (old B).
  • New: more texts, gap sentences, and rapid reading of short texts and graphs.
• Listening:
  • Removed: identifying a radio programme's section (old B) and the final phonetics part (old D).
  • Modified: drawings (old A), voicemails and announcements (old B), vox-pops (old C).
  • New: more audio documents, including a radio report.
• So any book or video teaching sentence reordering, reformulation or phonetics is pre-2023 material.

3. Scoring: what 434 means and how many correct answers it takes

• Official rule: each correct answer is 1 raw point; wrong and blank answers score 0 (livret 2026 p.26). The raw total is not the final score. The results page says it is "converti en un score calibré qui tient compte des différences de difficulté" of the questions (results FR). The method is not published. High confidence.

• CCI table for tests from 11 Dec 2023: CE and CO share one column (NCLC table). High confidence.

  NCLC	CE and CO score /699
  6	393-433
  7	434-461
  8	462-502
  9	503-545
  10	546-572

• IRCC:

  • For Express Entry, enter the "Équivalence ancien score" values. The /699 scores are not compatible with IRCC's system (IRCC).
  • CLB 7 starts at 249/360 for listening and 207/300 for reading.
  • IRCC also says that for tests taken from 11 Dec 2023 to 6 May 2024, CCI temporarily changed how NCLC levels were assigned so that they reflected raw scores, then reverted the change.
  • High confidence.

• Worked example from the sample attestation:

  • Reading: 459/699, equivalence 210/300, B2, NCLC 7.
  • Listening: 484/699, equivalence 289/360, B2, NCLC 8.
  • Source: sample attestation. It is an illustrative document. NCLC 7 sits inside the B2 band on it.

• Share of correct answers needed for 434. No official figure exists, so here are the anchors:

  • If the score were linear in raw answers, 434/699 = 62%, about 25/40.
  • The IRCC-column thresholds are 249/360 = 69% and 207/300 = 69%, about 28/40 if linear.
  • One third-party estimate puts NCLC 7 in listening at 22-26 raw, and the site itself calls this a planning estimate (OpenExamPrep CO).
  • My own estimate, by level: near-full marks on A1-A2 items, about 85% on B1 items, about 50% on B2 items and about 35% on C1 items (guessing included). This gives about 26/40.
  • Use 28/40 (70%) as the practice target and 30+ as comfortable. Confidence: low.

• Guessing: a 4-option guess is worth 0.25 point on average and a 3-option vox-pop guess 0.33. In CO you must click before the screen moves on, or the answer is blank (CO prep page).

4. Listening (CO): map by question position

The section boundaries and task types are official (CO sample p.2). The per-type counts are official except voicemails (CO prep page). That page lists 4 drawings, 4 announcements, 6 vox-pop questions, 2 chronicles, 6 interview questions, 1 report and 10 various documents. That is 33 questions with voicemails left out. Q5-20 holds 16 questions, so voicemails = 16 - 4 - 6 = 6 (inferred). The report carries 2 questions, so Q21-30 = 2 + 6 + 2 = 10. Positions inside a block follow the order of the sample and the tutorial (inferred, medium confidence).

The timings come from the CO sample and are the same in the 2025 tutorial: drawings, announcements, voicemails, chronicle, interviews, report, various.

The tutorial marks each screen with how many times its audio plays:

• Once: announcement, vox-pop, chronicle, report and various-document screens (announcement, vox-pop, chronicle, report, various).
• Twice: both interview screens (interview 1, interview 2).

Level by position:

• Questions 1-20 are short, everyday and A1-B1.
• From 21 on, the samples use C1 vocabulary: idioms of decline and doom, "a way to soothe our bad conscience", hedged conditionals.
• The level judgments are mine, from reading the CO sample. Official sources only say that the questions differ in difficulty.

5. Reading (CE): map by question position

All counts are official, from the section instructions in the CE sample. High confidence. The split inside 8-17 and 18-22 follows the sample's order.

The CE tutorial reuses these sample items (CE tutorial data). The section names match the third-party Section A-G labels used by one course provider (Alexa CE).

Reading-load note (my word counts from the sample): the sample documents run from about 50 to 300 words. Sixty minutes for 40 questions is about 90 seconds per question. The time pressure sits in 23-40, not in 1-22.

6. What recurs: counted

6.1 Questions per test form, by task type (official structure, repeated in every form)

Sources: CO prep page, CE sample.

Most repeated question goals across a CO form (my projection from the block counts):

• Purpose questions (why broadcast, why calling): about 10 (announcements and voicemails).
• Stance questions: 6.
• Picture matching: 4.
• The other 20 split between gist/opinion, detail and reason.

6.2 Question types in the 19 official CO sample items

Source: CO sample.

6.3 Question types in the 21 official CE sample items

Source: CE sample.

What this means: gap-fill is the single most repeated reading task (25% of every form). It tests vocabulary and collocation, not grammar. All 7 sample gap answers are content words chosen for meaning or collocation; none is a verb ending or an article.

7. Work-to-reward ranking and tactics

"Reward" = questions per form; every question is 1 raw point. The calibration may give harder items more weight, which is not published. "Effort" is my estimate for an English-speaking beginner over 8 weeks.

7.1 Listening, best ratio first

Tactics by type:

• Drawings (1-4).
  • Use the 5 s to spot what differs between the four drawings: place, object, action, number or time, weather, number of people.
  • Listen for that difference, not for the first noun you hear. Distractor drawings usually show something mentioned but not chosen.
  • Drill: describe 4 near-identical pictures aloud in French. Learn about 200 concrete words: shop, restaurant, transport, home, office, doctor, weather, clothes, plus numbers and clock times.
• Vox-pops (15-20).
  • Classify the speaker's final position. Ignore their reasons.
  • Totally for: no reservation.
  • For under conditions: "yes, but…", "on condition that…", "if…", "it depends", "in some cases", "up to a point".
  • Totally against: refusal words.
  • Since 2025 there is no "no opinion" option, so a hesitant speaker is almost always the conditional answer.
  • Drill: 3 short opinions on one topic per day; say the stance aloud in under 2 seconds.
• Voicemails (9-14).
  • Every message has this skeleton: greeting → who is calling → the reason → details (date, time, place) → call back.
  • The answer is in the clause after "I'm calling to / about / following…". Learn about 25 purpose verbs (section 10.2).
  • Contrast the close pairs: confirm / cancel / postpone / bring forward / modify; warn of / inform.
  • Use the 10 s before each message to read the 4 infinitives and translate them in your head.
• Announcements (5-8).
  • The first words name the setting (station, airport, shop, venue, town hall).
  • Then comes a change (delay, cancellation, platform or opening-hours change, closure) or an offer (sale, free event, sign-up).
  • The question is almost always the purpose. Learn the formulas in section 10.3.
• Interviews (23-28).
  • Read the 8 options in the 20 s and mark 1-2 key words in each.
  • First play: answer Q1, which is usually the guest's goal or main claim.
  • Second play: confirm, and answer Q2 (a reason or a detail).
  • The answer is a paraphrase, never the same words as the audio (section 9).
• Report (29-30). Expect a social or economic measure. Listen for the measure itself, who benefits, and the conditions or restrictions.
• Various documents (31-40).
  • Read the question stem: it names the speaker's act (announces, rejects, warns, denounces, demonstrates).
  • Learn the reporting and attitude verbs in section 10.8.
  • Decide whether the tone is positive, negative or neutral. Remove the extreme options, then always click an answer.
• Chronicles (21-22). The columnist's verdict on the concept is the target. Listen hardest to the last 15 seconds. Don't spend study time here until the items above are secure.

The official advice confirms the core habits:

• read the questions and options before the audio;
• answer even if unsure;
• train on key words, dates, places and numbers.

Source: CO prep page. Official free practice: CO sample plus its audio ZIP, and the CO tutorial. Workshop replays and a Facebook group are linked from the preparation page.

7.2 Reading, best ratio first

Tactics by type:

• Everyday documents (1-7).
  • Read the question first, then the title, headline, bold text and logo.
  • For "what is this document", learn the names of document types (section 10.6).
  • Aim for 1 minute each.
• Rapid reading (18-22).
  • Turn the target information into 2-3 French synonyms before scanning. In the sample, "sales job" matches an ad for a sales représentant.
  • For the graph item, translate the comment into a shape: "stagnation then rise" means flat, then going up.
• Gap sentences (8-13).
  • Read the whole sentence, then test each option as a phrase (a fire se déclare; you avez rendez-vous).
  • All four options are usually grammatical, so choose by meaning and collocation.
  • Keep a collocation notebook by theme: verb + noun pairs.
• Admin and professional documents (23-32).
  • For emails and letters, expect the genre questions: what the writer does (confirms, claims, informs, invites, reminds), plus one detail.
  • Read the numbers and conditions twice (prices, "included" vs "extra", deadlines, times).
  • Expect one opinion paragraph in this block: in the sample, the author's irony is the answer.
• Press articles (33-40).
  • Read the two questions, then skim the article. Answer the gist question from the title and the first and last sentences.
  • Then find the exact sentence for the detail question.
  • Give each article about 3.5 minutes, roughly 15 minutes for the block.
• Gap texts (14-17). Use logic, not knowledge:
  • Contrast markers ("from the most professional… to the most…") call for an antonym.
  • A list ("free, by subscription or pay-per-item") calls for a word that covers all the items.
  • Register: formal texts take formal adjectives.
• Order and time:
  • Navigation is free, so work in order but skip anything that takes more than 2 minutes.
  • The finish button appears only after every question has been seen (livret 2026 p.14).
  • Leave 3 minutes to fill every blank with a guess.

7.3 Learn-first order (the most repeated, cheapest items first)

1. Vox-pop stance markers (CO 15-20, 6 questions per form). Purpose verbs, which are shared by announcements, voicemails, writer's-purpose questions and "what this card lets you do": about 10 CO questions (5-14) plus several CE purpose questions (1-7, 23-32) at each sitting.
2. Everyday-scene words, numbers, times, dates and prices (CO 1-14; CE 1-7, 14-15, 20-21 in the samples).
3. Document-type words and graph-trend words (CE 1-7 and 18-22, 12 questions).
4. Collocations for the gap sentences (CE 8-13, 6 questions).
5. Formal email phrases (CE 23-32, 10 questions).
6. Reporting and attitude verbs, plus connectors (CO 21-40 and CE 23-40, 38 questions, but harder).

8. A points plan that reaches 28/40 in each test

How to read this:

• In listening, 18 of the 28 points come from questions 1-20. That needs about 90% accuracy on A1-B1 audio.
• The other 10 points come from questions 21-40. Guessing alone is worth about 5 there. Real understanding of roughly 5-7 of those 20 items, plus removing one option on the rest, closes the gap.
• In reading, 15 points come from the low-level blocks and 13 from the harder ones. This is more realistic than in listening, because text stays on screen and cognates help.
• These targets are a planning tool (my arithmetic), not an official conversion.

9. Common traps

9.1 Reading distractors, counted

I classified the 27 wrong options of the 9 official CE sample items whose full text is visible (Q4, Q14-Q21) (CE sample):

Rules:

• The right answer is a paraphrase.
• An option that copies text words is suspect.
• Watch the quantity words "only", "all", "always", "never", "mandatory".
• Check the tense of every verb in the option against the text.

False friends in the CE sample (4 found):

• location = rental
• prestation = service
• règlement = payment (it also means rules)
• assister à = attend

9.2 Listening traps

There are no official transcripts, so these come from the option sets plus third-party guides:

• Near-synonym purposes (official option sets): warn of a delay / announce a change / give news; inform a client / book an appointment / offer a discount. Learn exact verb meanings (CO sample).
• Stance flips in vox-pops: a speaker who starts positive and then adds a condition is "conditional", not "for" (FrenchForPR, third-party).
• Picture keyword trap: a heard word matches a drawing but the situation doesn't (FrenchForPR, third-party).
• Echo distractors: options built from numbers, dates, places or words actually heard (FrenchForPR, LanguageNext CO samples, third-party).
• Hedge vs extreme: an option states a firm position where the speaker hedged. Also sound-alike words and reversed cause and effect (OpenExamPrep CO, third-party).
• Process traps:
  • Missing the second interview question because you focused on the first.
  • Panicking after a miss in the fast 5-20 block and losing the next item too.
  • Fatigue in 31-40.
  • Training with the pause button, which the exam doesn't have.
  • Source for all four: FrenchForPR, third-party.
• Timing trap (official windows): the answer window after announcements and voicemails is only 10 s. Click a choice at once, then use the next pre-read window for the next item (CO sample).
• Material trap:
  • Pre-December-2023 books teach removed tasks (radio-programme sections, phonetics, reordering, synonym sentences) (2023 changes).
  • Pre-2025 sets train 4-option vox-pops and single-play interviews.
  • Several centre pages still describe 60/50-question papers.

10. Vocabulary domains and core word sets

10.1 Domains in the 40 official sample items (primary domain per item; my tagging)

Source: CO sample, CE sample.

Counting secondary tags too, work and professional life touches 7 items. Combined with the block names ("administrative and professional documents", "voicemails", "public announcements"), that makes it the most important domain for the cheap blocks. Numbers appear in the documents of 7 of the 21 CE items. In 3 of them (Q14, Q20, Q21), numbers are part of the answer or the distractors.

10.2 Purpose verbs (announcements, voicemails, writer's purpose)

annoncer, informer, prévenir, avertir, rappeler, confirmer, annuler, reporter, décaler, avancer, modifier, proposer, offrir, inviter, demander, réclamer, remercier, s'excuser, réserver, prendre rendez-vous, prendre des nouvelles, féliciter, convoquer, relancer, signaler, présenter, promouvoir.

10.3 Announcement formulas

Mesdames et Messieurs; votre attention s'il vous plaît; le train à destination de / en provenance de; voie, quai; retard de… minutes; supprimé / annulé; correspondance; embarquement immédiat, porte; nous informons notre aimable clientèle; le magasin fermera ses portes dans…; promotion, rayon, caisse; spectacle, séance, billetterie, entrée gratuite, inscription, horaires.

10.4 Voicemail formulas

c'est X de la part de Y; je vous appelle au sujet de / pour / suite à / concernant; je voulais vous prévenir que; votre commande est prête / arrivée; votre rendez-vous du… est reporté / annulé / avancé; pourriez-vous me rappeler au…; je vous laisse mon numéro; merci de confirmer.

10.5 Stance markers (vox-pops)

• For: je suis tout à fait pour; j'y suis favorable; c'est une excellente idée; sans hésiter; évidemment; je ne vois que des avantages.
• Conditional: oui, mais…; pourquoi pas, à condition que / de…; seulement si…; ça dépend (de)…; dans certains cas; en partie; à la rigueur; pas à n'importe quel prix; il faudrait d'abord…; je suis pour, sauf si…
• Against: je suis complètement contre; pas du tout; absolument pas; hors de question; je m'y oppose; je ne vois pas l'intérêt; ça ne sert à rien; c'est une mauvaise idée.

10.6 Document types (CE 1-7; "what is this document")

un article (de journal), une brochure / un dépliant, une affiche, une publicité, une petite annonce, une offre d'emploi, une invitation, un programme, un mode d'emploi / une notice, un règlement (intérieur), un formulaire, une facture, un reçu, un courriel, une lettre de réclamation, un avis, un sondage / une enquête.

10.7 Graph and statistics words (CE 22, CE 1-7 surveys)

augmenter / augmentation / hausse / progression / croissance / doubler / recrudescence / pic / record; baisser / baisse / diminution / chute / recul; stagner / stagnation / stable / se stabiliser / se maintenir / fluctuer; la majorité, la plupart, la moitié, un tiers, un quart, X % des personnes interrogées, par rapport à, depuis.

10.8 Reporting and attitude verbs (CO 31-40 stems; CE author-attitude items)

annoncer, affirmer, estimer, souligner, rappeler, regretter, déplorer, dénoncer, critiquer, ironiser, alerter, mettre en garde, s'inquiéter, se féliciter, saluer, approuver, soutenir, défendre, revendiquer, réclamer, contester, rejeter, s'opposer à, manifester, nuancer, relativiser.

The official samples alone use 10 of these in stems or options: annoncer, rejeter, alerter, dénoncer, manifester, regretter, ironiser, déplorer, affirmer, estimer.

10.9 Formal and administrative phrases (CE 23-32)

conformément à; à titre d'avance / à titre gratuit; en supplément; compris / inclus; ci-joint / veuillez trouver; merci de + infinitive; suite à votre demande; nous avons le plaisir / le regret de; dans les meilleurs délais; sous réserve de; à compter du; au plus tard le; la réservation, la facture, le devis, l'acompte, le remboursement, l'abonnement, la résiliation, les conditions générales de vente, débiter / créditer.

10.10 Connectors (stance flips, argument structure)

mais, pourtant, cependant, toutefois, en revanche, malgré, bien que, même si, alors que, car, puisque, en effet, d'ailleurs, donc, ainsi, à la différence de, dans un premier temps / aussitôt après.

The CE sample texts alone contain pourtant, malgré, mais, d'ailleurs, en effet, à la différence de, au contraire, aussitôt après and dans un premier temps (CE sample).

11. Grammar that the documents use most

Each feature is listed with where it appears in the official samples. Counts are from my reading of the CO sample and CE sample.

1. Infinitive of purpose.
   • pour / appeler pour / permettre de / avoir pour ambition de + infinitive.
   • 4 of the 19 CO items have stems built this way. Their 16 options are all infinitive phrases.
   • In CE, the card item ("lets you…") works the same way.
   • Highest priority: it drives about 10 CO questions per form.
2. Negation.
   • ne…pas plus ne…personne (the gap item about an empty room), ne…que and ne…aucun (the canteen points), and negated infinitives ("feel they don't control their life").
   • These carry meaning in gap sentences and in detail traps.
3. Tense and time contrast.
   • Passé composé vs future or a promise (promettre de + infinitive).
   • This is the whole point of one sample item's two distractors (Q21). Announcements also turn on changes of time.
4. Conditional for hedged or reported claims.
   • serait, seraient, sonneraient, faudrait in 2 CO opinion items; devrait in the science article; on croirait in the same article.
   • Needed for 31-40 and press items.
5. Relative pronouns, and en / y.
   • dont (a CO report stem), sur lesquelles, ce qui.
   • en five times in the CE texts: "there are many of them", "a burger is worth only 4", "gives none", "eat 100 of them", "makes them…".
6. Passive and pronominal passive.
   • In the CE texts and options: sont proposées, sont alloués, sont stockés, est constitué, a été créé, s'est déclaré and others, at least 10 in all. One of them is a gap answer.
   • The CO stems also use est diffusé.
7. Comparatives and quantities.
   • "three times more … than", "the most fertile", "healthier", "the most popular", percentages, "the majority".
   • Needed for surveys and graphs (CE Q4, Q13).
8. Present participle and gerund. pouvant (able to), considérant que (considering that), permettant (allowing), in press and admin texts.
9. Numbers, times and prices read and heard.
   • Examples: 8h, 20 heures, 21h00, 50,00 €, 3,50 €, 60 %, 4 000 points, 1,80 €, 60 000 €.
   • Numbers appear in the documents of 7 of the 21 CE items (Q4, Q5, Q6, Q14, Q15, Q20, Q21). In 3 of them (Q14, Q20, Q21), numbers are part of the answer or the distractors.

12. Where sources conflict

1. Interview plays.
   • The Oct 2024 sample booklet says every audio plays once (CO sample).
   • The tutorial dated 26-08-25 says both interviews play twice (interview 1, interview 2).
   • Third-party sites date the change to 1 September 2025 and say "certain" interviews play twice (tcftefprep changes, tcftefprep double listen). They cite a tefcanada.ca page that could not be reached from here.
   • I found no official news item about it. Train for one play; treat the second play as a bonus.
2. Vox-pop options. 4 options (with "no opinion") in the Oct 2024 booklet; 3 in the Aug 2025 tutorial (tutorial data). Third-party sites agree it is 3 since September 2025.
3. Score scale.
   • /699 (CCI table) vs /360 and /300 (the attestation's "Équivalence ancien score" column, which IRCC uses).
   • Many third-party pages mix them up or call the /360 figures "raw scores" (OpenExamPrep scoring, LanguageNext CE).
4. Scratch paper. Allowed according to the 2026 administrator booklet; forbidden according to the 2016 e-TEF manual and one third-party guide (OpenExamPrep CO). Ask your centre.
5. Going back in reading.
   • Official booklets: you navigate at your own pace. The 2016 manual: free navigation, answers editable until you click finish.
   • One third-party page says you cannot go back (LanguageNext CE). Another says you can go back and flag questions (OpenExamPrep CE).
   • The official version wins.
6. CO type counts. The official page lists 33 of the 40 questions by type and leaves out voicemails. Third-party pages copy the 33 and leave 7 unassigned (LanguageNext CO samples, OpenExamPrep CO). The block arithmetic gives 6 voicemails (section 4).
7. Listening answer windows. One third-party guide says 10-15 s to pre-read and 5-10 s to answer (OpenExamPrep CO). The official windows vary by type: 5-20 s before, 10-30 s after.
8. Reading block structure. One third-party guide's 4-part split (A-D) and its question counts don't match the official 5 blocks (LanguageNext CE).

13. Open questions

• The raw-to-score conversion and any per-item weighting by difficulty are not published.
• The exact positions of announcements, voicemails and vox-pops inside 5-20 are inferred, and so is the 3-interview split.
• Whether all interviews or only some play twice in live forms; the official tutorial shows both of its interviews played twice.
• Whether your centre provides scratch paper.
• Which accents the audio uses. Third-party sites mention France, Québec, Belgium, Switzerland and Africa (LanguageNext CO). Not verified officially.
• No candidate-recalled CO/CE topics for the current TEF Canada were found. The search budget ran out before Reddit, YouTube or Facebook compilations could be checked. A follow-up search could look for them.

Source list

Official (Le français des affaires, CCI Paris IDF)

• CO sample booklet: https://lefrancaisdesaffaires.fr/wp-content/uploads/2024/10/tef-exemples-epreuves-co.pdf
• CO audio ZIP: https://www.lefrancaisdesaffaires.fr/wp-content/uploads/2024/10/tef-bandes-sons-co.zip
• CE sample booklet: https://www.lefrancaisdesaffaires.fr/wp-content/uploads/2024/10/tef-exemples-epreuves-ce.pdf
• CO tutorial (title dated 26-08-25): https://www.lefrancaisdesaffaires.fr/documents/Tutoriel-TEF-CO/story.html
  • Screen data: https://www.lefrancaisdesaffaires.fr/documents/Tutoriel-TEF-CO/html5/data/js/data.js
  • Title and date: https://www.lefrancaisdesaffaires.fr/documents/Tutoriel-TEF-CO/html5/data/js/frame.js
  • Individual screens: .../5ou2BVFdAlr.js, 6oR49GejuWT.js, 69dexwxTr1c.js, 64ClGUae9JM.js, 623uZf8MQB9.js, 6FP67T2f09P.js, 5plXjiBsZK8.js, 5br04szXYuE.js, 6nmeCWFZ14X.js, 5wt73LpCi5w.js, 6BM7OtcwchK.js, 6XFcXGKfaRK.js, 6pAN6CCvcgY.js, 6km7hNQPrEM.js, 6rVmWb4g9LJ.js
• CE tutorial data: https://www.lefrancaisdesaffaires.fr/documents/Tutoriel_TEF_CE/html5/data/js/data.js
• CO preparation page (11 Dec 2023): https://www.lefrancaisdesaffaires.fr/en/prepare-yourself-for-the-oral-comprehension-test-of-the-tef/
• December 2023 changes: https://www.lefrancaisdesaffaires.fr/evolutions-tef-2023/
• TEF Canada presentation: https://www.lefrancaisdesaffaires.fr/en/candidate/test-evaluation-francais/tef-canada/presentation/
• TEF Canada preparation: https://www.lefrancaisdesaffaires.fr/en/candidate/test-evaluation-francais/tef-canada/preparation/
• Results pages: https://www.lefrancaisdesaffaires.fr/candidat/test-evaluation-francais/tef-canada/resultats/ and https://www.lefrancaisdesaffaires.fr/en/candidate/test-evaluation-francais/tef-canada/results/
• NCLC table: https://www.lefrancaisdesaffaires.fr/wp-content/uploads/2024/10/correspondance-tef-nclc.pdf
• Sample attestation: https://www.lefrancaisdesaffaires.fr/wp-content/uploads/2025/11/exemple-attestation-tefcanada.pdf
• Administrator booklets: https://www.lefrancaisdesaffaires.fr/wp-content/uploads/2026/04/2026-livret-du-responsable-tef.pdf and https://www.lefrancaisdesaffaires.fr/wp-content/uploads/2025/10/2025-livret-du-responsable-tef.pdf
• e-TEF candidate manual (2016, old format): https://ualberta.ca/fr/campus-saint-jean/media-library/programmes/edl/qmmanuele-tefcanadamars-20160-1.pdf

Government

• IRCC Express Entry language testing: https://www.canada.ca/en/immigration-refugees-citizenship/services/immigrate-canada/express-entry/documents/language-requirements/language-testing.html

Books and platforms

• Hachette FLE book: https://www.hachettefle.com/grands-ados-et-adultes/tef/tef-test-devaluation-de-francais-nouvelle-edition-livre-dentrainement
• ABC TEF listing: https://export.side.fr/en/9782090398083-abc-tef
• PrepMyTEF: https://www.prepmyfuture.com/products/tef

Third-party (tactics only)

• https://learnfrenchwithalexa.com/blog/tef-compréhension-orale
• https://learnfrenchwithalexa.com/blog/tef-compréhension-écrite
• https://learnfrenchwithalexa.com/important-information-regarding-changes-to-the-tef
• https://www.tcftefprep.com/blog/tef-canada-exam-format-changes-2026
• https://www.tcftefprep.com/blog/tef-canada-listening-section-f-double-listen
• https://www.languagenext.com/blog/tef-canada-listening/
• https://www.languagenext.com/blog/tef-canada-reading/
• https://www.languagenext.com/blog/tef-canada-listening-sample-papers/
• https://open-exam-prep.com/study-guides/tef-canada/listening-comprehension/listening-overview
• https://open-exam-prep.com/study-guides/tef-canada/reading-comprehension/reading-overview
• https://open-exam-prep.com/study-guides/tef-canada/exam-structure-nclc-scoring/nclc-cefr-scoring
• https://frenchforpr.com/tef-canada/comprehension-orale/
• https://www.digischool.fr/articles/fle/tef-format-epreuves-deroule/
• https://pvtistes.net/forum/entree-express-hors-quebec/121714-preparation-tef.html