MAI-Thinking-1
MAI-Thinking-1 is worth evaluating for coding, rag, and agents when its provider route and context window match the workload.
Use it for
- Teams evaluating coding, rag, and agents
- Workloads that can use a 256k context window
- Buyers comparing 1 tracked provider route
Do not use it for
- Vision or document-understanding workloads
- Family
- MAI
- Released
- 2026-06-02
- Context
- 256k
- Parameters
- 1T total / 35B active
- Architecture
- Mixture of Experts
- Specialization
- reasoning
- Openness
- Proprietary
- License
- ProprietaryCommercial use: conditional
- Weights
- Not released
- Code
- Unknown
- Training
- Pretrained
Cheapest of 1 route · Microsoft Foundry
About
MAI-Thinking-1 is Microsoft AI's flagship reasoning model, built from scratch on enterprise-grade commercially licensed data without third-party distillation. The sparse mixture-of-experts model activates about 35B parameters from roughly 1T total parameters, supports a 256K-token context window, and targets frontier reasoning and software engineering work at a mid-weight price point. Microsoft reports 97% on AIME 2025, 94.5% on AIME 2026, 84.2% on GPQA Diamond, 87.7% on LiveCodeBench v6, 73.5% on SWE-bench Verified, and 52.8% on SWE-bench Pro. In a 1,276-task Surge blind side-by-side evaluation, it narrowly beat Claude Sonnet 4.6 but trailed Claude Opus 4.6. It supports function calling and developer instructions through the Chat Completions API.
MAI-Thinking-1 is a proprietary model in the MAI family. The structured metadata tracks a 256k-token context window, reasoning, function calling, and tool use. This page tracks provider routes through Microsoft Foundry. Headline tracked benchmarks include AIME 2025 97.0, AIME 2026 94.5, and HMMT February 2026 84.9.
Top use-case fit: coding, agents, and build tasks
Coding
3 relevant benchmarks in the decision map.
RAG
Included by capability and metadata signals in the decision map.
Agents
2 relevant benchmarks in the decision map.
Provider price ladder
Compare API pricing across 1 providers for input and output tokens, batch, and cached reads when available.
| Provider | Input / 1M | Output / 1M | Route |
|---|---|---|---|
| Microsoft Foundry | - | - | ServerlessPartial |
Available via routers & gateways(5)
LiteLLM
GatewayOpen-source Python SDK and proxy server that unifies 100+ LLM APIs behind a single OpenAI-compatible interface, with load balancing, cost tracking, and configurable failover.
Portkey
GatewayProduction AI gateway routing to 1,600+ LLMs with failover, load balancing, semantic caching, and guardrails; Apache 2.0 core is fully self-hostable with the complete feature set.
Azure AI Foundry Model Router
RouterMicrosoft Azure AI Foundry's native model router that uses a trained ML model to route each prompt in real time to the optimal Azure-hosted model, with Balanced/Cost/Quality mode selection and automatic failover.
Helicone
GatewayObservability-first AI gateway with routing, caching, rate limiting, and request tracing; Apache 2.0 open-source core with a managed hosted tier for logging and analytics.
Kong AI Gateway
GatewayMulti-LLM AI gateway built on Kong Gateway 3.x, adding semantic routing, load balancing, guardrails, and MCP traffic analytics as plugins over Kong's existing API management platform.
Capabilities
Benchmark peer barsfor Coding
Benchmark scores(10)
| Benchmark | Score | Version | Source |
|---|---|---|---|
| AIME 2025 | 97.0 | AIME 2025 | https://microsoft.ai/news/introducing-mai-thinking-1/ |
| AIME 2026 | 94.5 | AIME 2026 | https://microsoft.ai/news/introducing-mai-thinking-1/ |
| HMMT February 2026 | 84.9 | HMMT Feb 2026 | https://microsoft.ai/wp-content/uploads/2026/06/main_20260602_2.pdf |
| Google-Proof Q&A | 84.2 | GPQA Diamond | https://microsoft.ai/wp-content/uploads/2026/06/main_20260602_2.pdf |
| LiveCodeBench | 87.7 | v6 | https://microsoft.ai/wp-content/uploads/2026/06/main_20260602_2.pdf |
| Terminal-Bench 2.0 | 46.0 | Terminal-Bench 2.0 | https://microsoft.ai/wp-content/uploads/2026/06/main_20260602_2.pdf |
| SWE-bench Verified | 73.5 | SWE-bench Verified | https://microsoft.ai/wp-content/uploads/2026/06/main_20260602_2.pdf |
| SWE-bench Pro | 52.8 | Public dataset | https://microsoft.ai/wp-content/uploads/2026/06/main_20260602_2.pdf |
| MMLU PRO | 85.0 | MMLU-Pro (accuracy) | https://llm-stats.com/benchmarks/mmlu-pro |
| MultiChallenge | 53.0 | Multi-Challenge leaderboard rank 15 of 28 (accuracy%) | https://llm-stats.com/benchmarks/multichallenge |
Migration checks
No linked migration route is available for this model yet.
Rankings & picks(1)
Frequently asked questions
What is the context window of MAI-Thinking-1?
MAI-Thinking-1 has a context window of 256k tokens.
When was MAI-Thinking-1 released?
MAI-Thinking-1 was released on 2026-06-02.
Which providers offer MAI-Thinking-1?
MAI-Thinking-1 is available from 1 provider: Microsoft Foundry.
What benchmarks has MAI-Thinking-1 been tested on?
MAI-Thinking-1 has been evaluated on 10 benchmarks, including AIME 2025, AIME 2026, HMMT February 2026, Google-Proof Q&A, LiveCodeBench.
Cheapest of 1 route · Microsoft Foundry