MAI-Thinking-1 vs Phi-4-reasoning
Change modelsMAI-Thinking-1Microsoft
Phi-4-reasoningMicrosoft
IndexAll categories, out of 100.
53.5 ±8.1 best
27.1 ±6.4
GapIndex points below the top.
Top
−26.4
CoverageIndex weight that has results.
70%
85% best
SpeedOutput tokens per second.
—
41/s
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2026-06
2025-04
Input / 1MUS dollars per 1M input tokens.
—
—
Output / 1MUS dollars per 1M output tokens.
—
—
1K chats2K tokens in, 500 out, each.
—
—
1K documents20K tokens in, 1K out, each.
—
—
1K agent tasks200K tokens in, 20K out, each.
—
—
ContextMaximum tokens in one request.
256K best
32K
AgenticMulti-step tasks with tools.
46.6 best
5.0
CodingCode writing and repair.
55.2 best
34.0
ReasoningLogic problems and puzzles.
—
22.1
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
53.0 best
33.3
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
68.5 best
21.1
MathMath problems.
52.6 best
35.5
Benchmarks
3 sharedOnly MAI-Thinking-1: 9Only Phi-4-reasoning: 9
MAI-Thinking-1
Phi-4-reasoning