HSK 6 Vocabulary with SRS: A 5000-Word Roadmap

5월 28, 2026

I spent four months grinding HSK 6 vocabulary while working full-time as a software engineer. My first attempt was a disaster—I made 2,000 cards with pinyin on the front, burned out after six weeks, and couldn't recognize 成语 (chéngyǔ, idioms) when I saw them in actual articles. My second attempt, using a structured SRS approach with tone-coded cards and separate recognition/production decks, got me to a 78% pass on the reading section.

HSK 6 requires roughly 5,000 words. That's not just memorizing definitions—it's distinguishing 制定 (zhìdìng, to formulate) from 制订 (zhìdìng, to draw up), recognizing tone shifts in compounds, and producing the right character when you hear 发展 (fāzhǎn) vs 发现 (fāxiàn). An SRS system handles the forgetting curve, but only if you design your cards correctly.

TL;DR
HSK 6 needs ~5,000 words mastered for recognition and 2,000-3,000 for production. Use separate decks: hanzi-only for recognition, pinyin+audio for production. Tone-code your cards (red=1st, yellow=2nd, green=3rd, blue=4th). Budget 25-30 new cards/day for 6 months with 40-60 min/day reviews. Watch for cognate confusion (近义词), tone shifts in compounds, and measure-word traps.

Why HSK 6 Vocabulary Breaks Most SRS Workflows

HSK 6 isn't just "more words." The jump from HSK 5 (2,500 words) to HSK 6 (5,000 words) introduces:

  • Abstract vocabulary: 抽象 (chōuxiàng, abstract), 概括 (gàikuò, to summarize), 推理 (tuīlǐ, reasoning)
  • Near-synonyms with subtle distinctions: 尽管 (jǐnguǎn) vs 虽然 (suīrán) vs 即使 (jíshǐ)—all mean "although" but with different nuances
  • Four-character idioms (成语): 画蛇添足 (huà shé tiān zú), 守株待兔 (shǒu zhū dài tù)
  • Formal/written register: 鉴于 (jiànyú, in view of), 倘若 (tǎngruò, if)

Most learners create cards like:

Front: 制定
Back: to formulate (policy, plan)

This works until you encounter 制订 (same pinyin, same tone, nearly identical meaning). Or until you see 制定 in a sentence and can't recall it without the pinyin cue.

The Two-Deck System: Recognition vs Production

I split my HSK 6 vocabulary into two decks after realizing I could "recognize" 4,000 words but only "produce" 1,500. Recognition means seeing 经济发展 (jīngjì fāzhǎn, economic development) and understanding it. Production means hearing "economic development" and writing/saying 经济发展 with correct tones.

Recognition Deck (4,000-5,000 cards)

Front: 制定 (hanzi only, no pinyin)
Back: to formulate (policy, plan) • zhìdìng • 政府制定了新的环保政策

Key principles:

  • Hanzi-only front forces character recognition without pinyin crutches
  • Example sentence shows the word in context (critical for HSK 6 reading comprehension)
  • Tone-coded hanzi using color: 制 (blue for 4th tone) 定 (blue for 4th tone)

I used Pleco's built-in tone colors initially, then migrated to SmartRecall where I could batch-apply tone coding to 2,000+ cards at once.

Production Deck (2,000-3,000 cards)

Front: to formulate (policy, plan) • [audio: zhìdìng]
Back: 制定

Key principles:

  • English + audio on the front (no hanzi)
  • Hanzi-only back to test production
  • Smaller deck because you don't need to produce every word—focus on high-frequency and exam-relevant terms

For HSK 6, I prioritized production for:

  • Top 2,000 frequency words
  • All 成语 (idioms) from official lists
  • Verbs and adjectives (nouns are easier to recognize than produce)

Tone-Coded Cards: Why Color Matters

Mandarin has four tones plus neutral tone. HSK 6 loves minimal pairs:

  • 发展 (fāzhǎn, development) vs 发现 (fāxiàn, to discover)
  • 质量 (zhìliàng, quality) vs 质量 (zhíliàng, mass—yes, same hanzi, different tones)

I tone-code using:

  • Red: 1st tone (high, flat)
  • Yellow: 2nd tone (rising)
  • Green: 3rd tone (dipping)
  • Blue: 4th tone (falling)
  • Gray: neutral tone

Example: 经济 becomes 经(red)济(blue) = jīngjì.

This visual cue trains your brain to associate the character with its tone before you even hear it. After 3,000 reviews, I could "see" the tone when reading—my internal voice automatically used the correct tone.

SmartRecall's bulk tone-coding feature saved me 20+ hours. I imported a CSV with hanzi and pinyin, and it auto-applied colors based on tone marks.

The 6-Month Roadmap for Working Learners

Assuming you're at HSK 5 level (2,500 words) and studying 60 minutes/day:

Month 1-2: Foundation (25 new cards/day)

  • Week 1-4: Add 700 recognition cards (top 1,000 HSK 6 frequency words minus HSK 5 overlap)
  • Week 5-8: Add 700 more recognition cards + start 300 production cards for highest-frequency verbs
  • Daily load: 25 new + 40-60 reviews = 50-70 min

Focus on single-character and two-character words first. Avoid 成语 until you have a base of 1,500 HSK 6 words.

Month 3-4: Expansion (30 new cards/day)

  • Week 9-12: Add 840 recognition cards (mid-frequency HSK 6 words)
  • Week 13-16: Add 840 recognition cards + 400 production cards (common adjectives and nouns)
  • Daily load: 30 new + 80-100 reviews = 60-80 min

Introduce 成语 here—10-15 per week. Create separate cards for each idiom:

  • Recognition: 画蛇添足 → to ruin something by adding unnecessary details (lit. "draw a snake and add feet")
  • Production: to ruin by over-embellishing → 画蛇添足

Month 5-6: Consolidation (20 new cards/day)

  • Week 17-20: Add 560 recognition cards (low-frequency but exam-relevant words)
  • Week 21-24: Add 560 recognition cards + 500 production cards (remaining 成语 and formal vocabulary)
  • Daily load: 20 new + 100-120 reviews = 60-90 min

By month 6, you're reviewing 120+ cards/day. This is sustainable if you use FSRS (Free Spaced Repetition Scheduler) or SM-2 algorithms—both adjust intervals based on your actual performance.

Total after 6 months:

  • 4,200 recognition cards
  • 1,200 production cards
  • ~5,400 total cards (accounting for overlap and deletions)

Common Pitfalls and How to Avoid Them

1. Cognate Confusion (近义词)

HSK 6 tests near-synonyms relentlessly:

  • 尽管 (jǐnguǎn, although—factual concession)
  • 虽然 (suīrán, although—general)
  • 即使 (jíshǐ, even if—hypothetical)

Solution: Create comparison cards.

Front: 尽管 vs 虽然 vs 即使
Back: 尽管=factual ("although it rained, we went") • 虽然=general ("although he's busy") • 即使=hypothetical ("even if it rains")

I made 50 comparison cards for the most confusing sets. SmartRecall's tagging system let me group them under "near-synonyms" for targeted review before the exam.

2. Tone Shifts in Compounds

Single characters change tone in compounds:

  • 一 (yī, one) → 一个 (yí ge, one [measure word])—tone shifts to 2nd
  • 不 (bù, not) → 不是 (bú shì, is not)—tone shifts to 2nd before 4th tone

Solution: Always include audio on production cards. Don't rely on pinyin alone—your brain will default to the citation tone.

3. Measure Word Overload

HSK 6 expects you to know 50+ measure words:

  • 一条河 (yì tiáo hé, one river—条 for long, narrow things)
  • 一座桥 (yí zuò qiáo, one bridge—座 for large structures)
  • 一首诗 (yì shǒu shī, one poem—首 for songs/poems)

Solution: Create dedicated measure-word cards with multiple examples.

Front: 座 (zuò)
Back: MW for buildings, mountains, bridges • 一座山 • 一座城市 • 一座桥

4. Passive Review Trap

Clicking "Good" on a card you half-remembered is a trap. HSK 6 reading passages move fast—you need instant recognition.

Solution: Use strict grading. If you hesitated more than 2 seconds on a recognition card, mark it "Again." For production cards, if you got the tone wrong, mark it "Again."

I kept my retention rate at 85-88% (not 90%+) to ensure I was actually learning, not just maintaining easy cards.

Measuring Progress Beyond Card Counts

Card counts are vanity metrics. What matters:

Reading Speed Test (Monthly)

  • Take a 500-character HSK 6 reading passage
  • Time yourself
  • Count unknown words

Target progression:

  • Month 1: 8-10 min, 15-20 unknown words
  • Month 3: 5-6 min, 8-12 unknown words
  • Month 6: 3-4 min, 3-5 unknown words

Listening Comprehension (Bi-weekly)

  • Listen to a 3-minute HSK 6 audio clip (no transcript)
  • Write down key points in English
  • Check against transcript

Target: 70%+ comprehension by month 4.

Production Test (Monthly)

  • Write a 200-character essay on a random HSK 6 topic (环保, 科技, 教育)
  • Use only words from your production deck
  • Check for tone/character errors

Target: 150+ characters with <5 errors by month 5.

Tools and Resources

SRS Apps:

  • SmartRecall: Best for bulk imports, tone-coding, and FSRS scheduling. I used it for 80% of my cards.
  • Pleco: Great for on-the-go review and built-in dictionary lookups.
  • Anki: Free and powerful, but tone-coding requires manual CSS or add-ons.

Vocabulary Sources:

  • Official HSK 6 word list (5,000 words)—download from Hanban or Chinese Testing International
  • "HSK 6 Vocabulary" by Sinolingua (book with example sentences)
  • Du Chinese app (graded readers at HSK 6 level for context)

Audio:

  • Forvo.com for native speaker pronunciations
  • Google Translate's audio (surprisingly accurate for Mandarin)

The Final Month: Exam-Specific Drills

Four weeks before your HSK 6 exam, shift from pure SRS to exam simulation:

  • Week -4: Take a full practice test. Identify weak vocabulary categories (成语, formal verbs, abstract nouns).
  • Week -3: Create 100 "high-stakes" cards from your weak categories. Review these daily in addition to regular SRS.
  • Week -2: Reduce new cards to 10/day. Focus on retention and speed.
  • Week -1: No new cards. Review only. Take one final practice test.

I added 150 "exam-specific" cards in my final month—words that appeared in 3+ practice tests but weren't in my deck yet. SmartRecall's "cram mode" let me review these separately without disrupting my main deck's scheduling.

What 5,000 Words Actually Feels Like

After six months and 12,000+ reviews, I could:

  • Read 人民日报 (People's Daily) articles with 90%+ comprehension
  • Watch Chinese news without subtitles and catch 70% of content
  • Write emails to Chinese colleagues without constant dictionary lookups

But I still struggled with:

  • Regional dialects and slang (HSK 6 is standard Mandarin only)
  • Handwriting characters from memory (typing is easier)
  • Speaking fluently at native speed (production ≠ conversation)

HSK 6 vocabulary is a foundation, not a finish line. But it's the foundation that unlocks authentic content—novels, podcasts, professional documents.

Your First 100 Cards

Start tonight:

  1. Download the official HSK 6 word list
  2. Extract the top 100 frequency words you don't know
  3. Create 100 recognition cards (hanzi front, English + example back)
  4. Add tone colors manually or use SmartRecall's auto-coding
  5. Review 20 cards tomorrow, add 25 new ones

Six months from now, you'll have 5,000 words and a passing HSK 6 score. The system works—but only if you start.

Alex Chen

Alex Chen