# Kimi K2.5 vs Phi 3 Mini 4k Instruct

> Kimi K2.5 is the stronger model overall, scoring 48.1 to 27.9 on the Noometry Index.

- Canonical page: https://noometry.com/compare/kimi-k2-5-vs-phi-3-mini-4k-instruct
- Last updated: 2026-10-11
- Shared benchmarks: 18

## Summary

- They share 18 benchmarks with published results for both. Kimi K2.5 scores higher in 8 categories and Phi 3 Mini 4k Instruct in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Kimi K2.5 leads 65.1 to 27.6.
- The biggest single-benchmark swing is Chess Puzzles: 12% for Kimi K2.5 and 0% for Phi 3 Mini 4k Instruct.

## Snapshot

| | Kimi K2.5 | Phi 3 Mini 4k Instruct |
|---|---|---|
| Provider | Moonshot AI | Microsoft |
| Noometry Index | 48.1 | 27.9 |
| Rank | 57 | 328 |
| Context | 262K | — |
| Input $/M | $0.45 | — |
| Output $/M | $2.25 | — |
| Weights | Open | Open |

## Coding

- Kimi K2.5: 48.8 (#53)
- Phi 3 Mini 4k Instruct: 26.6 (#323)

| Benchmark | Kimi K2.5 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Coding | 1474 | 1093 |
| SWE-bench Verified | 73.8% | — |
| SWE-bench Verified (bash only) | 70.8% | — |
| LMArena WebDev | 1437 | — |
| SWE-bench Multilingual | 67.3% | — |
| SciCode | 49% | — |
| WeirdML | 45.6% | — |
| LiveBench Coding | — | 15.5% |
| ALE-Bench | 821.65 | — |
| HumanEval+ | — | 59.1% |
| MBPP+ | — | 54.2% |

## Agentic & Tool Use

- Kimi K2.5: 34.2 (#48)
- Phi 3 Mini 4k Instruct: —

| Benchmark | Kimi K2.5 | Phi 3 Mini 4k Instruct |
|---|---|---|
| Terminal-Bench | 43.2% | — |
| OSWorld | 63.3% | — |
| Vending-Bench 2 | 1,198 | — |

## Reasoning

- Kimi K2.5: 31.2 (#80)
- Phi 3 Mini 4k Instruct: 14.1 (#328)

| Benchmark | Kimi K2.5 | Phi 3 Mini 4k Instruct |
|---|---|---|
| Chess Puzzles | 12% | 0% |
| LMArena Hard Prompts | 1453 | 1072 |
| ARC-AGI-2 | 11.8% | — |
| SimpleBench | 46.8% | — |
| Kagi LLM Benchmark | 78.5% | — |
| NYT Connections (extended) | 69.9% | — |
| ARC-AGI-1 | 65.3% | — |
| CritPt | 3.1% | — |
| EnigmaEval | 3.4% | — |
| Thematic Generalization | 69.4% | — |
| LiveBench Reasoning | — | 26.8% |
| LiveBench Data Analysis | — | 34.7% |
| Adversarial NLI | — | 52.8% |
| BIG-Bench Hard | — | 71.7% |
| Epoch Capabilities Index | 148.03 | — |
| HellaSwag | — | 76.7% |
| LiveBench | — | 22.4% |
| WinoGrande | — | 70.8% |

## Math

- Kimi K2.5: 51.8 (#53)
- Phi 3 Mini 4k Instruct: 26.6 (#257)

| Benchmark | Kimi K2.5 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Math | 1470 | 1111 |
| MathArena Final-Answer Competitions | 62.3% | — |
| OTIS Mock AIME 2024-2025 | 92.2% | — |
| LiveBench Math | — | 15.7% |
| FrontierMath (Feb 2025 set) | 27.9% | — |
| FrontierMath Tier 4 (v1) | 4.2% | — |

## Knowledge

- Kimi K2.5: 53.6 (#56)
- Phi 3 Mini 4k Instruct: 28.5 (#246)

| Benchmark | Kimi K2.5 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Expert | 1466 | 1045 |
| GPQA Diamond | 87.6% | — |
| Humanity's Last Exam | 24.4% | — |
| SimpleQA Verified | 34.3% | — |
| Vectara Hallucination Rate | 14.2% | — |
| ARC (AI2) Challenge | — | 84.9% |
| MMLU | — | 68.8% |
| OpenBookQA | — | 88% |
| TriviaQA | — | 64% |

## Multimodal

- Kimi K2.5: 41.1 (#39)
- Phi 3 Mini 4k Instruct: —

| Benchmark | Kimi K2.5 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Vision | 1269 | — |
| LMArena Document | 1430 | — |

## Multilingual

- Kimi K2.5: 53.9 (#53)
- Phi 3 Mini 4k Instruct: 26.3 (#280)

| Benchmark | Kimi K2.5 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Non-English | 1433 | 1021 |
| LMArena Chinese | 1495 | 1021 |
| LMArena French | 1454 | 1076 |
| LMArena German | 1441 | 1044 |
| LMArena Japanese | 1421 | 935 |
| LMArena Korean | 1410 | 905 |
| LMArena Russian | 1435 | 1022 |
| LMArena Spanish | 1450 | 1085 |

## Instruction Following

- Kimi K2.5: 75.3 (#64)
- Phi 3 Mini 4k Instruct: 47.7 (#303)

| Benchmark | Kimi K2.5 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Instruction Following | 1431 | 1053 |
| LiveBench Instruction Following | — | 39.1% |

## Long Context

- Kimi K2.5: 52.1 (#7)
- Phi 3 Mini 4k Instruct: 31.7 (#276)

| Benchmark | Kimi K2.5 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Longer Query | 1445 | 1044 |
| Fiction.LiveBench | 86.1% | — |
| CL-bench | 19.3% | — |
| CL-bench Life | 13.2% | — |

## Writing & Preference

- Kimi K2.5: 65.1 (#53)
- Phi 3 Mini 4k Instruct: 27.6 (#300)

| Benchmark | Kimi K2.5 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Text | 1445 | 1073 |
| LMArena Creative Writing | 1423 | 1037 |
| LMArena Multi-Turn | 1444 | 1018 |
| EQ-Bench Creative Writing | 1579 | — |
| LiveBench Language | — | 9.2% |

## FAQ

### Is Kimi K2.5 better than Phi 3 Mini 4k Instruct?

Kimi K2.5 is the stronger model overall, scoring 48.1 to 27.9 on the Noometry Index.

### Is Kimi K2.5 or Phi 3 Mini 4k Instruct better for coding?

Kimi K2.5 scores higher on coding benchmarks: 48.8 versus 26.6 in the Noometry coding category.

### How many benchmarks do Kimi K2.5 and Phi 3 Mini 4k Instruct share?

18 benchmarks have published results for both models. Kimi K2.5 has 51 scored results on Noometry and Phi 3 Mini 4k Instruct has 35.