Leaderboard / Microsoft
Phi-4
Microsoft OPEN MIT
Verdict
Phi-4 is behind the leaders: #295 of 406 by SIscope score. Even across the board, with no weak spot in what's been tested. Cheaper than 97% of priced models. Weights are open, so you can run it yourself.
SIscope score
26
#295 of 406
Human-vote rank
#317
of 419 · score 1217
Price per 1M tokens, blended
$0.087
$0.07 in · $0.14 out
Context window
16K
16,384 tokens
Shape
Each spoke is a category, 0 at the centre, 100 at the edge.
Category scores
Percentiles: 90 means better than about 90% of tested models.
Human-vote score over time
Text arena, recent updates.
Independent test runs
Run by Epoch AI on the same harness for every model. Bars show the share of questions answered correctly.
GPQA Diamond56%
PhD-level science multiple choice questions
Better than 30% of 333 tested models. Run Jan 31, 2025.
OTIS Mock AIME 2024-2514%
Competition maths problems in the style of AIME
Better than 17% of 313 tested models. Run Feb 25, 2025.
Human-vote arenas
People compare two anonymous answers and pick the better one. Scores work like chess ratings.
| Arena | Rank | Score | Range (95%) | Votes | Updated |
|---|---|---|---|---|---|
| Text arena Everyday chat, blind human votes | #317 of 419 | 1217 | 1212–1221 | 24,126 | Oct 28, 2026 |
| WebDev arena Building web apps, blind human votes | Not on this board yet | ||||
| Vision arena Questions about images, blind human votes | Not on this board yet | ||||
| Agent arena Real multi-step agent sessions | Not on this board yet | ||||
Source: Arena (formerly LMArena) leaderboard dataset, CC BY 4.0. Fetched Nov 4, 2026.
Price
Input
$0.07
per 1M tokens
Output
$0.14
per 1M tokens
Blended
$0.087
3 in : 1 out, per 1M
Context window
16K
max tokens per request
Speed
Coming soon
our own tests
List price from the OpenRouter models API (microsoft/phi-4), Nov 4, 2026. Providers can charge differently.