AILIMIT.WATCH
LIVE

Mistral Le Chat

Last verified: Jun 8, 2026

8 sourced changes·100% official·verified against official docs, changelogs & Wayback·last checked Jun 8, 2026

01 / VERDICT

Caution

Frequent recent tightening — high constraint volatility right now.

WHY CAUTION — trend: tightening · drift -16% over 90d · 2 tightenings in 90d

Constraint Stability

80/100

higher = fewer recent restrictions

tightening-16% / 90d

Mistral Le Chat has become more restrictive over time, mainly in model availability.

Most noticeable: Broad disruption — integrations on Mixtral 8x7B etc. forced to migrate within ~4 months

Recency
29
days since it last tightened
Velocity
1.6
tightenings per year
Silent
0%
of tightenings were unannounced
Reversals
50%
of tightenings later reversed (same dimension)

Mistral Le Chat constraint metrics — computed from 4 tracked tightenings · median 303 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
Exposed to usage-based cost changes across the team.
high
Startup / cost-scaling
Exposed to pricing volatility and usage-based cost as usage scales.
high
Enterprise / long-term dependency
Continuity exposure: pricing-predictability, ownership / policy shifts.
high

02 / ANALYSIS

Cheap to start, expensive to stay pinned — Mistral Le Chat pulls in two directions at once. The generosity is real on price: September 2024 cut Mistral Large API pricing 33% ($3/$9 to $2/$6) and Small by 80%, and added a no-card 'Experiment' free tier for prototyping. Working against that is a brisk deprecation cadence. February 2024 handed the tiny/small/medium alias endpoints three-month deprecations with no auto-redirect, forcing code changes; November 2024 then deprecated 9+ endpoints in one batch — Mistral 7B, Mixtral 8x7B/8x22B and others — retiring them over Mar–Jun 2025 and compressing broad migrations into about four months. The February 2025 Le Chat Pro launch ($14.99/mo) drew explicit ceilings too: a ~25 msg/day free cap, with Mistral Large and Flash Answers behind Pro. And lately the tilt is upward — June 2026's OCR 4 doubled the base rate from $2 to $4 per 1,000 pages, with the mistral-ocr-latest alias silently shifting callers onto the pricier model, while mistral-small-latest raised input 50% and output 100%. Read together: aggressive new-model pricing, but model longevity treated as disposable and periodic upward resets. Adopt it cheaply if you like, but plan for forced migrations on a roughly quarterly-to-annual rhythm, pin to versioned model names rather than -latest aliases (which can move you to a costlier model with no code change), and re-verify per-token and per-page rates given the 2026 upward drift on OCR and Small.

03 / CURRENT LIMITS

04 / WHERE IT MATTERS

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

Agentic / high-volume costhigh exposure
Jun 2026↓Base OCR rate doubled ($2 → $4 / 1000 pages); callers pinned to mistral-ocr-latest pay 2x as the alias moved to OCR 4.·source
Jun 2026↓mistral-small-latest pricing raised: input +50% ($0.1 → $0.15), output doubled ($0.3 → $0.6) per M tokens.·source
Nov 2024↓Broad disruption — integrations on Mixtral 8x7B etc. forced to migrate within ~4 months·source
Sep 2024↑Large across-the-board API cost cut + zero-cost prototyping tier·source
Specific-model dependencelow exposure
Feb 2025→Free users get an explicit daily ceiling and lose Mistral Large / Flash Answers·source
Nov 2024↓Broad disruption — integrations on Mixtral 8x7B etc. forced to migrate within ~4 months·source
Feb 2024↓API users with hard-coded model strings had to change code across integrations·source
High-frequency usagelow exposure
Feb 2025→Free users get an explicit daily ceiling and lose Mistral Large / Flash Answers·source
Sep 2024↑Large across-the-board API cost cut + zero-cost prototyping tier·source

05 / WHAT CHANGED

Technical 0 · Product 4 · Business 4 · Policy 0

  1. Nov 2024

    ↓ Broad disruption — integrations on Mixtral 8x7B etc. forced to migrate within ~4 months

    productImpact High · source
  2. Sep 2024

    ↑ Large across-the-board API cost cut + zero-cost prototyping tier

    businessImpact Medium · source
  3. Feb 2024

    ↓ API users with hard-coded model strings had to change code across integrations

    productImpact Medium · source
▸ View full constraint history (6 events)
  1. Jun 2026

    ↓ Base OCR rate doubled ($2 → $4 / 1000 pages); callers pinned to mistral-ocr-latest pay 2x as the alias moved to OCR 4.

    businessImpact Medium · source · secondary source

    ↔ shipped alongside Adds a new document intelligence model to the available toolset. (Jun 2026) — more for more, not a pure cut.

  2. Jun 2026

    ↓ mistral-small-latest pricing raised: input +50% ($0.1 → $0.15), output doubled ($0.3 → $0.6) per M tokens.

    businessImpact Medium · source · secondary source

    ↔ shipped alongside Adds a new document intelligence model to the available toolset. (Jun 2026) — more for more, not a pure cut.

  3. Feb 2025

    → Free users get an explicit daily ceiling and lose Mistral Large / Flash Answers

    businessImpact Medium · source · secondary source
  4. Nov 2024

    ↓ Broad disruption — integrations on Mixtral 8x7B etc. forced to migrate within ~4 months

    productImpact High · source
  5. Sep 2024

    ↑ Large across-the-board API cost cut + zero-cost prototyping tier

    businessImpact Medium · source
  6. Feb 2024

    ↓ API users with hard-coded model strings had to change code across integrations

    productImpact Medium · source
▸ Other product updates (2) — feature launches, not constraint changes
  1. Jul 2026

    ↑ Adds a new model capability for embodied navigation.

    productImpact Low · source · secondary source
  2. Jun 2026

    ↑ Adds a new document intelligence model to the available toolset.

    productImpact Low · source · secondary source

06 / OUTLOOK

Active

Active. Last 12 months: 2 pricing, 0 limit/quota, 2 model-availability changes (2 tightenings).

Observed past cadence — not a prediction.

07 / USER SIGNALS

What users report

  • • “The 'unlimited' messaging comes with a soft cap of ~150 messages/day under fair use.” source

08 / ALTERNATIVES

ALTERNATIVES

By constraint history — not capability.

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