Skip to main content
AI Choice EngineAI Choice Engine

Model comparison

Microsoft MAI-Thinking-1 vs Qwen 3.8 Max

Microsoft against Qwen, 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

vs

Qwen

Qwen 3.8 Max

Frontier · Open weights

AI model capability comparison
SpecificationMicrosoft MAI-Thinking-1Qwen 3.8 Max
ProviderMicrosoftQwen
TierFrontierFrontier
Context window256KWinner: 991K
Max output64KWinner: 131K
Input / 1M tokens$2$2
Output / 1M tokens$8Winner: $6
WeightsClosedOpen
Parameters~1T total / 35B active (sparse MoE)2.4T total / 95B active (MoE)
Reasoning levelslow, medium, highlow, high, max
Modalitiestexttext, image, video
API model idmai-thinking-1qwen3.8-max
ReleasedJune 2, 2026August 3, 2026
Artificial Analysis Intelligence Index (2026-08-14)55Winner: 58
AIME 2025 (2026-08-12)97Not verifiedUnverified
SWE-bench Verified (2026-06)73.5Winner: 85.6
GPQA Diamond (2026-06)84.2Winner: 92.6
Terminal-Bench 2.0 (2026-06)46Not verifiedUnverified
Terminal-Bench 2.1 (2026-08-03)Not verifiedUnverified86.6
Humanity's Last Exam (2026-08-14)Not verifiedUnverified56.2

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

  • 2026-09-01: vals.ai SWE-bench Verified leaderboard (independent, mini-swe-agent bash-only); not published by Qwen (official used SWE-bench Pro 67.7)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.
  • 2026-08-14: Z.ai GLM-5.3 blog (independent Z.ai-run, with tools; Qwen official no-tools 43.6 — both preserved)Frontier-knowledge ceiling: extremely hard multi-domain questions written to be near-impossible without deep expertise. Comparability: comparable with caveat — Subset/tool configurations (text-only vs with-tools) differ between vendors and materially change scores; must match configuration to compare.
  • 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.
  • 2026-08-03: Qwen official blog (vendor-run table)Agentic terminal work: multi-step tasks executed in a sandboxed shell environment. Comparability: comparable with caveat — Scores across Terminal-Bench major versions (v1 vs v2) are NOT comparable; harness configuration (container, time limits) also shifts results.
  • 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.

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. · Qwen 3.8 Max: $2/$6 per MTok (Alibaba Cloud Model Studio). Weights announced open, verify they have dropped.

FrontierRecord checked September 5, 2026

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.

FrontierOpen weightsRecord checked September 3, 2026

Qwen 3.8 Max

Qwen 3.8 Max is Alibaba's flagship — 2.4T MoE with 1M-class context, near frontier on the Intelligence Index.

Best for

  • Open-weight frontier work
  • Long-context
  • Multimodal

Watch out

Open weights dropped 2026-08-12 under a custom (non-Apache) licence with vision and 1M-context stripped from the open checkpoint — the open checkpoint is not the full API model. Verify the licence before commercial use.

When the cheaper one wins

Qwen 3.8 Max is cheaper on output at $6 per million tokens against $8 for Microsoft MAI-Thinking-1 — about 1.3×. 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 picker

Evidence 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.
  • 5/5 core specs verified on both sides — All core specifications verified for both models.
  • 3 shared named benchmarks with differing scores — Measured on: Artificial Analysis Intelligence Index, SWE-bench Verified, 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.

Related comparisons

Diving deeper on one model? Microsoft MAI-Thinking-1 · Qwen 3.8 Max

Common questions

Microsoft MAI-Thinking-1 vs Qwen 3.8 Max

Answered from the verified figures on this page rather than general guidance.

Is Microsoft MAI-Thinking-1 or Qwen 3.8 Max cheaper for input?

Both cost $2 per million input tokens at standard rates, so input price is not a deciding factor between them.

Is Microsoft MAI-Thinking-1 or Qwen 3.8 Max cheaper for output?

Qwen 3.8 Max is cheaper at $6 per million output tokens, against $8 for Microsoft MAI-Thinking-1 — roughly 1.3× 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. Qwen 3.8 Max has tiered pricing: $2/$6 per MTok (Alibaba Cloud Model Studio). Weights announced open, verify they have dropped.

Which has the larger context window, Microsoft MAI-Thinking-1 or Qwen 3.8 Max?

Qwen 3.8 Max accepts 991K tokens against 256K for Microsoft MAI-Thinking-1. This only matters if you routinely send very long documents or large codebases.

Do Microsoft MAI-Thinking-1 and Qwen 3.8 Max support the same reasoning levels?

Microsoft MAI-Thinking-1 exposes low, medium, high, while Qwen 3.8 Max exposes low, high, max.

Should I use Microsoft MAI-Thinking-1 or Qwen 3.8 Max?

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); Qwen 3.8 Max suits open-weight frontier work.

Can I self-host Microsoft MAI-Thinking-1 or Qwen 3.8 Max?

Qwen 3.8 Max 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.

  • Context: Qwen 3.8 Max takes 991K against 256K for Microsoft MAI-Thinking-1 — only decisive if your prompts approach the smaller window.
  • Measured capability: Qwen 3.8 Max leads Artificial Analysis Intelligence Index 58 to 55 (measured 2026-08-14).
  • Deployment: Qwen 3.8 Max publishes weights you can self-host; the other is API-only.
Run the model pickerEstimate stack costAll model comparisons

Choosing a harness rather than a model — Cursor, Copilot, Claude Code, Windsurf, or Lovable? Compare agentic harnesses · Latest releases.