[Verification Conflict] Mistral Open Models - Mistral Small 3

[github-actions]

Verification Conflict

Automated verification found conflicting information about:

Platform: Mistral Open Models
Feature: Mistral Small 3

Current Stored Data

What our data file currently says

{
  "name": "Mistral Small 3",
  "category": "other",
  "status": "ga",
  "gating": "free",
  "availability": [
    {
      "plan": "Self-hosted",
      "available": "✅",
      "limits": "Your hardware",
      "notes": "24B parameters"
    }
  ],
  "platforms": [
    {
      "platform": "Windows",
      "available": "✅",
      "notes": "Ollama, LM Studio"
    },
    {
      "platform": "macOS",
      "available": "✅",
      "notes": "Runs on 32GB RAM MacBook"
    },
    {
      "platform": "Linux",
      "available": "✅",
      "notes": "Single RTX 4090"
    },
    {
      "platform": "iOS",
      "available": "❌",
      "notes": ""
    },
    {
      "platform": "Android",
      "available": "❌",
      "notes": ""
    },
    {
      "platform": "Chrome",
      "available": "❌",
      "notes": ""
    },
    {
      "platform": "web",
      "available": "✅",
      "notes": "Web UIs"
    },
    {
      "platform": "terminal",
      "available": "✅",
      "notes": "Ollama CLI"
    },
    {
      "platform": "API",
      "available": "✅",
      "notes": "Local API"
    }
  ]
}

Model Responses

Gemini Flash

Result: Detected change

Full response

Based on the search results, here's an updated verification of your data for "Mistral Small 3":

1. Pricing Tier Availability:

• Subscription Plans: Self-hosted. It is also available on la Plateforme as mistral-small-latest or mistral-small-2501.
• Usage Limits: Your hardware.
• Match: CORRECT.

2. Platform/Surface Availability:

• Availability:
  • Windows: ✅ (via Ollama, LM Studio)
  • macOS: ✅ (Runs on 32GB RAM MacBook)
  • Linux: ✅ (Single RTX 4090)
  • iOS: ❌
  • Android: ❌
  • Chrome: ❌
  • Web: ✅ (Web UIs)
  • Terminal: ✅ (Ollama CLI)
  • API: ✅ (Local API)
• Match: CORRECT.

3. Current Status:

• Status: GA (Generally Available).
• Match: CORRECT.

4. Access Gating:

• Gating: Free (Apache 2.0 license).
• Match: CORRECT.

5. Regional Availability:

• Regional Availability: No restrictions (Apache 2.0 license).
• Match: CORRECT.

6. Official URL:

• While a direct product page isn't readily available, the following URLs are relevant:
  • Mistral AI's main website: N/A
  • Hugging Face model page: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503
• The URLs are active and accessible.

7. Recent Changes:

• Updates:
  • Mistral Small 3.1: Announced in March 2025, featuring improved text performance, multimodal understanding, and an expanded context window.
  • Mistral Small 3.2: Released in June 2025, focusing on improvements to instruction following, output stability, and function calling robustness.
  • Mistral 3 Family: In December 2025, Mistral AI launched the Mistral 3 family, including smaller models (Ministral 3) designed for edge AI applications.
• Given the release of newer versions (3.1 and 3.2) and the Mistral 3 family, it's important to specify which version your database is referencing.

Summary of Findings:

Your stored data is mostly CORRECT, but requires an update to reflect the latest versions and related changes:

• The original Mistral Small 3 has been superseded by Mistral Small 3.1 and Mistral Small 3.2. You should specify which version your database entry refers to.
• Consider adding entries for the Ministral 3 family if you are tracking smaller, edge-optimized models.
• The core details (pricing, platforms, status, gating, regional availability) remain largely consistent.

Perplexity

Result: No change detected

Full response

1. Pricing tier availability

Stored data is CORRECT.
Mistral Small 3 is available via Self-hosted plan under Apache 2.0 license, with limits determined by user hardware (e.g., single RTX 4090 or 32GB RAM MacBook).[1][2][4]

2. Platform/surface availability

Stored data is CORRECT.

• Windows: ✅ (via Ollama, LM Studio).[4]
• macOS: ✅ (runs on 32GB RAM MacBook).[1][2][4]
• Linux: ✅ (single RTX 4090).[2][4]
• iOS: ❌ (no mentions of native support).[1][2][4]
• Android: ❌ (no mentions of native support).[1][2][4]
• Chrome: ❌ (no browser extension or native support noted).[1][2][4]
• web: ✅ (via Web UIs for local deployment).[1]
• terminal: ✅ (Ollama CLI).[4]
• API: ✅ (local API).[1][4]

3. Current status

Stored data is CORRECT.
Model is GA (generally available, released Jan 30, 2025, with active downloads and deployments).[1][4]

4. Access gating

Stored data is CORRECT.
It is free (Apache 2.0 license, open weights for self-hosting, no paid or invite requirements).[1][2]

5. Regional availability

Stored data is CORRECT.
No restrictions (Apache 2.0 license enables global use; no country-specific limitations mentioned).[1][2]

6. Official URL

Stored data lacks explicit URL; official page is https://mistral.ai/news/mistral-small-3, which is active and details the model.[1]

7. Recent changes

Stored data requires update: New version Mistral Small 3.1 released Mar 17, 2025 (improved text/multimodal performance, 128k context; still 24B params, Apache 2.0, same hardware reqs).[2]
No other changes to original Small 3 in last 30 days (post-Feb 18, 2026). Original remains available.[1][2][3]

Action Required

Please manually verify this feature's current status:

1. Check the official pricing page
2. Verify platform availability
3. Update the data file if needed
4. Close this issue with findings

---

This issue was automatically generated by the pricing verification system.

[snapsynapse]

Resolved in PR #399