Model comparison
Microsoft MAI-Thinking-1 vs Mistral Large 3
Microsoft against Mistral, compared on context, price, and verified benchmark results.
Catalog record checked September 5, 2026Individual provider fields may changeEvidence confidence: High — see receipts below
Microsoft
Microsoft MAI-Thinking-1
Frontier
Mistral
Mistral Large 3
Frontier · Open weights
| Specification | Microsoft MAI-Thinking-1 | Mistral Large 3 |
|---|---|---|
| Provider | Microsoft | Mistral |
| Tier | Frontier | Frontier |
| Context window | 256K | Winner: 262K |
| Max output | 64K | Winner: 262K |
| Input / 1M tokens | $2 | Winner: $0.50 |
| Output / 1M tokens | $8 | Winner: $1.50 |
| Weights | Closed | Open |
| Parameters | ~1T total / 35B active (sparse MoE) | 675B total / 41B active (sparse MoE) |
| Reasoning levels | low, medium, high | Not verifiedUnverified |
| Modalities | text | text, image |
| License | Not disclosedUnverified | Apache 2.0 |
| API model id | mai-thinking-1 | mistral-large-3 |
| Released | June 2, 2026 | December 1, 2025 |
| Artificial Analysis Intelligence Index (2026-08-14) | Winner: 55 | 46 |
| AIME 2025 (2026-08-12) | 97 | Not verifiedUnverified |
| SWE-bench Verified (2026-06) | 73.5 | Not verifiedUnverified |
| GPQA Diamond (2026-06) | Winner: 84.2 | 43.9 |
| Terminal-Bench 2.0 (2026-06) | 46 | Not verifiedUnverified |
Prices are USD per million tokens at standard rates, excluding batch and caching discounts. Bold indicates the better figure where one is objectively better. Values we could not confirm from the provider are shown as “Not verified” rather than estimated.
Benchmark receipts
- 2026-08-14: Artificial AnalysisComposite index blending reasoning, knowledge, and coding evals into one 0–100 score. Comparability: directly comparable — AA occasionally rebaselines the index scale between snapshots — a score captured on one date is only comparable to same-snapshot scores (check measuredAt).
- 2026-08-12: Microsoft AI — Introducing MAI-Thinking-1 (vendor, 256k output)Competition mathematics: AIME 2025 exam problems, typically pass@1 with tools disallowed or single-attempt code execution per the harness. Comparability: comparable with caveat — 30-question total denominator: one problem = ~3.3 points. Tool-use policy (calculator/code interpreter) must match between compared models.
- 2026-06: Microsoft AI model page (vendor, as labeled: Terminal-Bench 2.0 — NOT comparable to 2.1 observations)Real GitHub issue resolution: does the model's patch pass the hidden tests. Comparability: comparable with caveat — Post-audit vendor claims and pre-audit scores sit on different task trust levels; scaffolding (agent harness, compute budget) also dominates results. Never aggregate across scaffolds.
- 2025-12-02: Mistral — Introducing Mistral 3 (vendor, base model, 5-shot no CoT); independent review concursGraduate-level science reasoning: multiple-choice questions written by domain PhDs. Comparability: comparable with caveat — Small question pool: differences under ~2 points are within run-to-run noise. Chain-of-thought vs direct answering must match to compare.
Pricing tiers: Microsoft MAI-Thinking-1: Global Standard $2.00/$8.00 per MTok, cached input $0.20 (Azure Retail Prices API, effective 2026-08-01, uniform across ~22 public regions; US Gov regions carry premiums). Still preview; Global Standard only, no PTU yet. 256K context with a 64K output cap that includes reasoning tokens. ~1T/35B sparse MoE; reasoning is always on (no effort levels). Microsoft's first true reasoning model, trained without OpenAI distillation. · Mistral Large 3: Mistral first-party API $0.50/$1.50 per MTok; cached input $0.05/MTok. Open weights (Apache 2.0), EU-hosted.
Microsoft MAI-Thinking-1
MAI-Thinking-1 is Microsoft's first in-house reasoning model, reducing dependence on OpenAI.
Best for
- Enterprise (Azure)
- Reasoning
- Microsoft ecosystem
Watch out
Foundry-gated preview; no MAI-2 exists yet. Of the Build 2026 MAI wave, only MAI-Thinking-1 and MAI-Cyber-1-Flash ($0.60/$3.50, released 2026-07-27) have published token meters.
Mistral Large 3
Mistral Large 3 is Mistral's flagship — 262K-context, multimodal, Apache-2.0 open-weight at $0.50/$1.50.
Best for
- Open-weight deployments
- Multimodal EU-hosted inference
- Balanced frontier work
Watch out
Verify multimodal support and EU data-residency on your endpoint.
When the cheaper one wins
Mistral Large 3 is cheaper on output at $1.50 per million tokens against $8 for Microsoft MAI-Thinking-1 — about 5.3×. Use the cheaper tier for classification, extraction, summarisation, and any task where the expensive model’s extra score does not change the accepted output. The expensive one only pays if your hardest task actually fails on the cheap tier. These are standard-tier API rates, excluding batch and cache discounts.
Run the model pickerEvidence confidence: High
How strong and complete the evidence behind this comparison is — not a prediction of which model is better.
- Pricing verified on both sides — Input and output rates are verified for both models.
- 4/5 core specs verified on both sides — Not published for at least one side: reasoning levels.
- 2 shared named benchmarks with differing scores — Measured on: Artificial Analysis Intelligence Index, GPQA Diamond.
- Verified within the last 90 days — Newest catalog check was 6 days ago.
- Both models carry source citations — Each side has at least two catalog sources on record.
Source receipts
Each catalog figure was checked against the provider or an independent second source on the date shown.
- Microsoft MAI-Thinking-1: Azure Retail Prices API (MAI-Thinking-1 $2/$0.20 cached/$8) (accessed 2026-09-04)
- Microsoft MAI-Thinking-1: AzureSpeed — AI model pricing tracker (same figures) (accessed 2026-09-04)
- Microsoft MAI-Thinking-1: Microsoft — Introducing MAI-Thinking-1 (~1T total / 35B active MoE) (accessed 2026-09-05)
- Microsoft MAI-Thinking-1: Microsoft Foundry — MAI-Thinking-1 usage (256K context, 64K output cap) (accessed 2026-09-05)
- Mistral Large 3: Mistral docs — mistral-large-2512 (675B total / 41B active, 256K) (accessed 2026-09-05)
- Mistral Large 3: Mistral — Introducing Mistral 3 (sparse MoE parameters) (accessed 2026-09-05)
Related comparisons
- Claude Fable 5.1 vs Microsoft MAI-Thinking-1
- Claude Fable 5.1 vs Mistral Large 3
- Claude Fable 5 vs Microsoft MAI-Thinking-1
- Claude Fable 5 vs Mistral Large 3
- Claude Mythos 5.1 vs Microsoft MAI-Thinking-1
- Claude Mythos 5.1 vs Mistral Large 3
Diving deeper on one model? Microsoft MAI-Thinking-1 · Mistral Large 3
Common questions
Microsoft MAI-Thinking-1 vs Mistral Large 3
Answered from the verified figures on this page rather than general guidance.
Is Microsoft MAI-Thinking-1 or Mistral Large 3 cheaper for input?
Mistral Large 3 is cheaper at $0.50 per million input tokens, against $2 for Microsoft MAI-Thinking-1 — roughly 4.0× the price. Output tokens usually dominate a real bill, so weigh the output rate more heavily than the input rate. These are base rates; Microsoft MAI-Thinking-1 has tiered pricing: Global Standard $2.00/$8.00 per MTok, cached input $0.20 (Azure Retail Prices API, effective 2026-08-01, uniform across ~22 public regions; US Gov regions carry premiums). Still preview; Global Standard only, no PTU yet. 256K context with a 64K output cap that includes reasoning tokens. ~1T/35B sparse MoE; reasoning is always on (no effort levels). Microsoft's first true reasoning model, trained without OpenAI distillation. Mistral Large 3 has tiered pricing: Mistral first-party API $0.50/$1.50 per MTok; cached input $0.05/MTok. Open weights (Apache 2.0), EU-hosted.
Is Microsoft MAI-Thinking-1 or Mistral Large 3 cheaper for output?
Mistral Large 3 is cheaper at $1.50 per million output tokens, against $8 for Microsoft MAI-Thinking-1 — roughly 5.3× the price. Output tokens usually dominate a real bill, so weigh the output rate more heavily than the input rate. These are base rates; Microsoft MAI-Thinking-1 has tiered pricing: Global Standard $2.00/$8.00 per MTok, cached input $0.20 (Azure Retail Prices API, effective 2026-08-01, uniform across ~22 public regions; US Gov regions carry premiums). Still preview; Global Standard only, no PTU yet. 256K context with a 64K output cap that includes reasoning tokens. ~1T/35B sparse MoE; reasoning is always on (no effort levels). Microsoft's first true reasoning model, trained without OpenAI distillation. Mistral Large 3 has tiered pricing: Mistral first-party API $0.50/$1.50 per MTok; cached input $0.05/MTok. Open weights (Apache 2.0), EU-hosted.
Which has the larger context window, Microsoft MAI-Thinking-1 or Mistral Large 3?
Mistral Large 3 accepts 262K tokens against 256K for Microsoft MAI-Thinking-1. This only matters if you routinely send very long documents or large codebases.
Should I use Microsoft MAI-Thinking-1 or Mistral Large 3?
Both sit in the frontier tier, so the choice usually comes down to price and context rather than capability. Microsoft MAI-Thinking-1 suits enterprise (azure); Mistral Large 3 suits open-weight deployments.
Can I self-host Microsoft MAI-Thinking-1 or Mistral Large 3?
Mistral Large 3 publishes open weights, but self-hosting and third-party provider access depend on the licence, hardware, serving support, and availability. Microsoft MAI-Thinking-1 is a closed model whose supported access paths are controlled by its provider.
Next step
Choosing between them
The verified figures that separate this pair, computed from the catalog rather than restated boilerplate.
- Input price: Mistral Large 3 costs $0.50 per 1M tokens versus $2 for Microsoft MAI-Thinking-1 — a 4x difference at the headline tier.
- Context: Mistral Large 3 takes 262K against 256K for Microsoft MAI-Thinking-1 — only decisive if your prompts approach the smaller window.
- Measured capability: Microsoft MAI-Thinking-1 leads Artificial Analysis Intelligence Index 55 to 46 (measured 2026-08-14).
- Deployment: Mistral Large 3 publishes weights you can self-host; the other is API-only.
Choosing a harness rather than a model — Cursor, Copilot, Claude Code, Windsurf, or Lovable? Compare agentic harnesses · Latest releases.