AILIMIT.WATCH
LIVE

DeepSeek

Last verified: Jul 22, 2026

26 sourced changes·15% official·verified against official docs, changelogs & Wayback·last checked Jul 22, 2026

01 / VERDICT

Caution

Frequent recent tightening — high constraint volatility right now.

WHY CAUTION — trend: unstable · drift -67% over 90d · 10 tightenings in 90d

Constraint Stability

28/100

higher = fewer recent restrictions

unstable-67% / 90d

DeepSeek has shown frequent recent tightening, mainly in pricing.

Most noticeable: DeepSeek V4 Flash input price $0.0938 → $0.0980 per 1M tokens (+4%)

Recency
1
days since it last tightened
Velocity
7.8
tightenings per year
Silent
0%
of tightenings were unannounced
Reversals
69%
of tightenings later reversed (same dimension)

DeepSeek constraint metrics — computed from 13 tracked tightenings · median 10 days between tightenings. Original to this site; not published by the vendor.

IMPACT BY PERSONA
Solo / occasional
Brief, occasional use is rarely touched by limit changes.
low
Power user / daily heavy
Daily-limit, session and rate constraints have been steady.
low
Team / shared workflow
Shared quotas, model availability and per-seat cost have held steady.
low
Startup / cost-scaling
Exposed to pricing volatility as usage scales.
high
Enterprise / long-term dependency
Continuity exposure: pricing-predictability, ownership / policy shifts.
high

02 / ANALYSIS

Cheap and volatile in equal measure — and the volatility touches nearly every dimension. The context window alone has moved both ways: halved from 128K to 64K in late 2024, restored to 128K in August 2025, then reset by V4 to a 1M default in April 2026. Model access churns just as hard. V3.2-Exp arrived in September 2025 with prices cut over 50%, V3.1-Terminus was retired weeks later after a brief comparison window, and the long-standing deepseek-chat and deepseek-reasoner endpoints are scheduled to go fully inaccessible after July 2026, forcing migration to V4. A regulatory availability risk sits on top: in June 2025 Berlin's data-protection authority reported the app to Apple and Google over China data transfers, pushing for removal in Germany. The V4-Flash price series shows the churn vividly — through mid-2026 the input and output rates oscillate almost weekly, drifting down toward $0.077/$0.15 per million but reversing direction repeatedly along the way. A user report names the trade directly: low per-token cost bought at the price of a lineup that keeps changing, each turn carrying a deprecation date. DeepSeek rewards the cost-sensitive and punishes the set-and-forget. Pin nothing without tracking its retirement schedule, expect prices to wobble in both directions, and account for the possibility that regional availability isn't guaranteed.

03 / CURRENT LIMITS

04 / WHERE IT MATTERS

Which usage patterns the recent changes affect — each line backed by the events driving it.

Specific-model dependencemedium exposure
Apr 2026↓Legacy API model endpoints scheduled for removal·source
Oct 2025↓Prior model retired; must move to V3.2-Exp·source
Long-context workloadslow exposure
Apr 2026↑New V4 models; default context raised to 1M·source
Aug 2025↑Context window restored (64K → 128K)·source
Nov 2024↓Main chat model's context window halved (128K → 64K)·source
Agentic / high-volume costlow exposure
Jul 2026↓DeepSeek V4 Flash input price $0.0938 → $0.0980 per 1M tokens (+4%)·source
Jul 2026↓DeepSeek V4 Flash output price $0.19 → $0.20 per 1M tokens (+4%)·source
Jul 2026↑DeepSeek V4 Flash input price $0.0980 → $0.0938 per 1M tokens (-4%)·source
Jul 2026↑DeepSeek V4 Flash output price $0.20 → $0.19 per 1M tokens (-4%)·source
Jul 2026↓DeepSeek V4 Flash input price $0.0770 → $0.0900 per 1M tokens (+17%)·source
Jul 2026↓DeepSeek V4 Flash output price $0.15 → $0.18 per 1M tokens (+17%)·source
Jul 2026↑DeepSeek V4 Flash input price $0.0900 → $0.0770 per 1M tokens (-14%)·source
Jul 2026↑DeepSeek V4 Flash output price $0.18 → $0.15 per 1M tokens (-14%)·source
Jul 2026↓DeepSeek V4 Flash input price $0.0890 → $0.0900 per 1M tokens (+1%)·source
Jul 2026↑DeepSeek V4 Flash input price $0.0980 → $0.0890 per 1M tokens (-9%)·source
Jul 2026↑DeepSeek V4 Flash output price $0.20 → $0.18 per 1M tokens (-8%)·source
Jul 2026↓DeepSeek V4 Flash input price $0.0900 → $0.0980 per 1M tokens (+9%)·source
Jul 2026↓DeepSeek V4 Flash output price $0.18 → $0.20 per 1M tokens (+9%)·source
Jun 2026↓DeepSeek V4 Flash input price $0.0900 → $0.0980 per 1M tokens (+9%)·source
Jun 2026↓DeepSeek V4 Flash output price $0.18 → $0.20 per 1M tokens (+9%)·source
Jun 2026↑DeepSeek V4 Flash input price $0.0980 → $0.0900 per 1M tokens (-8%)·source
Jun 2026↑DeepSeek V4 Flash output price $0.20 → $0.18 per 1M tokens (-8%)·source
Jun 2026↑DeepSeek V4 Flash input price $0.0983 → $0.0980 per 1M tokens (+0%)·source
Jun 2026↑DeepSeek V4 Flash output price $0.20 → $0.20 per 1M tokens (+0%)·source
Sep 2025↑API token cost cut ~50%·source

05 / WHAT CHANGED

Technical 2 · Product 3 · Business 20 · Policy 1

  1. Aug 2025

    ↑ Context window restored (64K → 128K)

    technicalImpact Low · source · secondary source
  2. Jun 2025

    ↓ Regulatory pressure to pull app in Germany (availability risk)

    policyImpact Medium · source
  3. Nov 2024

    ↓ Main chat model's context window halved (128K → 64K)

    technicalImpact Medium · source · secondary source
▸ View full constraint history (26 events)
  1. Jul 2026

    ↓ DeepSeek V4 Flash input price $0.0938 → $0.0980 per 1M tokens (+4%)

    businessImpact Medium · source
  2. Jul 2026

    ↓ DeepSeek V4 Flash output price $0.19 → $0.20 per 1M tokens (+4%)

    businessImpact Medium · source
  3. Jul 2026

    ↑ DeepSeek V4 Flash input price $0.0980 → $0.0938 per 1M tokens (-4%)

    businessImpact Medium · source
  4. Jul 2026

    ↑ DeepSeek V4 Flash output price $0.20 → $0.19 per 1M tokens (-4%)

    businessImpact Medium · source
  5. Jul 2026

    ↓ DeepSeek V4 Flash input price $0.0770 → $0.0900 per 1M tokens (+17%)

    businessImpact Medium · source
  6. Jul 2026

    ↓ DeepSeek V4 Flash output price $0.15 → $0.18 per 1M tokens (+17%)

    businessImpact Medium · source
  7. Jul 2026

    ↑ DeepSeek V4 Flash input price $0.0900 → $0.0770 per 1M tokens (-14%)

    businessImpact Medium · source
  8. Jul 2026

    ↑ DeepSeek V4 Flash output price $0.18 → $0.15 per 1M tokens (-14%)

    businessImpact Medium · source
  9. Jul 2026

    ↓ DeepSeek V4 Flash input price $0.0890 → $0.0900 per 1M tokens (+1%)

    businessImpact Medium · source
  10. Jul 2026

    ↑ DeepSeek V4 Flash input price $0.0980 → $0.0890 per 1M tokens (-9%)

    businessImpact Medium · source
  11. Jul 2026

    ↑ DeepSeek V4 Flash output price $0.20 → $0.18 per 1M tokens (-8%)

    businessImpact Medium · source
  12. Jul 2026

    ↓ DeepSeek V4 Flash input price $0.0900 → $0.0980 per 1M tokens (+9%)

    businessImpact Medium · source
  13. Jul 2026

    ↓ DeepSeek V4 Flash output price $0.18 → $0.20 per 1M tokens (+9%)

    businessImpact Medium · source
  14. Jun 2026

    ↓ DeepSeek V4 Flash input price $0.0900 → $0.0980 per 1M tokens (+9%)

    businessImpact Medium · source
  15. Jun 2026

    ↓ DeepSeek V4 Flash output price $0.18 → $0.20 per 1M tokens (+9%)

    businessImpact Medium · source
  16. Jun 2026

    ↑ DeepSeek V4 Flash input price $0.0980 → $0.0900 per 1M tokens (-8%)

    businessImpact Medium · source
  17. Jun 2026

    ↑ DeepSeek V4 Flash output price $0.20 → $0.18 per 1M tokens (-8%)

    businessImpact Medium · source
  18. Jun 2026

    ↑ DeepSeek V4 Flash input price $0.0983 → $0.0980 per 1M tokens (+0%)

    businessImpact Medium · source
  19. Jun 2026

    ↑ DeepSeek V4 Flash output price $0.20 → $0.20 per 1M tokens (+0%)

    businessImpact Medium · source
  20. Apr 2026

    ↑ New V4 models; default context raised to 1M

    productImpact Low · source
  21. Apr 2026

    ↓ Legacy API model endpoints scheduled for removal

    productImpact Medium · source

    ↔ shipped alongside New V4 models; default context raised to 1M (Apr 2026) — more for more, not a pure cut.

  22. Oct 2025

    ↓ Prior model retired; must move to V3.2-Exp

    productImpact Low · source
  23. Sep 2025

    ↑ API token cost cut ~50%

    businessImpact Low · source
  24. Aug 2025

    ↑ Context window restored (64K → 128K)

    technicalImpact Low · source · secondary source
  25. Jun 2025

    ↓ Regulatory pressure to pull app in Germany (availability risk)

    policyImpact Medium · source
  26. Nov 2024

    ↓ Main chat model's context window halved (128K → 64K)

    technicalImpact Medium · source · secondary source

06 / OUTLOOK

Volatile

Volatile. Last 12 months: 20 pricing, 1 limit/quota, 3 model-availability changes (11 tightenings).

Observed past cadence — not a prediction.

07 / USER SIGNALS

USER SIGNALS — HACKER NEWS · GITHUB · APP STORE

25 complaint mentions · last 180d

25 HN

  • Pricing·2080%
  • Performance·312%
  • Limits / quota·28%

Hacker News, GitHub issues & Apple App Store reviews matching complaint keywords — not a scientific sample. See methodology.

08 / ALTERNATIVES

ALTERNATIVES

By constraint history — not capability.

More stable:Bolt.new, Leonardo.ai, Ideogram
Similar risk:GitHub Copilot, Tabnine, ElevenLabs, Mistral Le Chat
→ Full comparison matrix

09 / RELATED