Model comparison
Kimi K3 vs Phi-4
Kimi K3 is the stronger model overall, scoring 59.5 to 31.2 on the Noometry Index. Phi-4 costs 69× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Last verified . 21 shared benchmarks.
MicrosoftSummary
- They share 21 benchmarks with published results for both. Kimi K3 scores higher in 9 categories and Phi-4 in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K3 leads 74.2 to 20.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 97.2% for Kimi K3 and 13.8% for Phi-4.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $3 / $15 for Kimi K3.
- Kimi K3 accepts more context: 1.05M tokens versus 128K.
Side by side
| Kimi K3 | Phi-4 | |
|---|---|---|
| Provider | Moonshot AI | Microsoft |
| Noometry Index | 59.5 | 31.2 |
| Released | 2026-07-16 | 2024-12-11 |
| Weights | Open | Open |
| Context window | 1.05M | 128K |
| Max output | 1.05M | 4K |
| Input $ / M tokens | $3 | $0.07 |
| Output $ / M tokens | $15 | $0.14 |
| Results tracked | 53 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
Kimi K3: 61.0 (#10), Phi-4: 34.4 (#239)
| Benchmark | Kimi K3 | Phi-4 |
|---|---|---|
| LMArena Coding | 1508 | 1231 |
| DeepSWE | 68.5% | — |
| FrontierCode | 44.2% | — |
| LMArena WebDev | 1654 | — |
| FrontierSWE | 25.9% | — |
| SciCode | 59.5% | — |
| WeirdML | 82.6% | — |
| BigCodeBench Instruct | — | 45.5% |
| LiveBench Coding | — | 30.7% |
| BigCodeBench Complete | — | 55.4% |
| ALE-Bench | 1,524 | — |
Agentic & Tool Use Kimi K3 leads
Kimi K3: 41.8 (#20), Phi-4: 22.8 (#128)
| Benchmark | Kimi K3 | Phi-4 |
|---|---|---|
| APEX-Agents | 50.6% | — |
| Berkeley Function Calling Leaderboard | — | 28.8% |
| τ²-bench Banking | 37.1% | — |
| PostTrainBench | 32% | — |
| BALROG | — | 11.6% |
| GBAEval | 48.3% | — |
| GDP.pdf | 19% | — |
| Vending-Bench 2 | 5,165 | — |
Reasoning Kimi K3 leads
Kimi K3: 63.0 (#17), Phi-4: 17.7 (#291)
| Benchmark | Kimi K3 | Phi-4 |
|---|---|---|
| Chess Puzzles | 39% | 1% |
| LMArena Hard Prompts | 1496 | 1220 |
| Epoch Capabilities Index | 157.45 | 130.42 |
| ARC-AGI-2 | 60.4% | — |
| SimpleBench | 60.7% | — |
| NYT Connections (extended) | 93.6% | — |
| ARC-AGI-1 | 94.5% | — |
| CritPt | 23.4% | — |
| LiveBench Reasoning | — | 47.8% |
| Mystery Game Puzzles | 26% | — |
| DTBench | 91.2% | — |
| LiveBench Data Analysis | — | 45.2% |
| LMCA | 52.7% | — |
| Surface Evolver Bench | 95% | — |
| ForecastBench | 61.1 | — |
| LiveBench | — | 41.6% |
Math Kimi K3 leads
Kimi K3: 74.2 (#16), Phi-4: 20.8 (#285)
| Benchmark | Kimi K3 | Phi-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 97.2% | 13.8% |
| LMArena Math | 1491 | 1246 |
| FrontierMath (Tiers 1-3) | 72.2% | — |
| FrontierMath Tier 4 | 39% | — |
| MathArena Final-Answer Competitions | 87.8% | — |
| ProofBench | 87% | — |
| LiveBench Math | — | 42% |
| MATH Level 5 | — | 64.9% |
Knowledge Kimi K3 leads
Kimi K3: 63.2 (#21), Phi-4: 32.6 (#209)
| Benchmark | Kimi K3 | Phi-4 |
|---|---|---|
| GPQA Diamond | 93.1% | 56.1% |
| LMArena Expert | 1521 | 1203 |
| SimpleQA Verified | 50.6% | — |
| Confabulations | — | 29.4% |
| Vectara Hallucination Rate | — | 3.7% |
| MMLU | — | 84.8% |
Multimodal Not comparable
Kimi K3: 37.8 (#70), Phi-4: —
| Benchmark | Kimi K3 | Phi-4 |
|---|---|---|
| Blueprint-Bench 2 | 29.5% | — |
| Furniture Assembly | 34.2% | — |
Multilingual Kimi K3 leads
Kimi K3: 56.3 (#21), Phi-4: 37.2 (#237)
| Benchmark | Kimi K3 | Phi-4 |
|---|---|---|
| LMArena Non-English | 1466 | 1197 |
| LMArena Chinese | 1529 | 1212 |
| LMArena French | 1491 | 1224 |
| LMArena German | 1488 | 1222 |
| LMArena Japanese | 1487 | 1158 |
| LMArena Korean | 1458 | 1151 |
| LMArena Russian | 1482 | 1209 |
| LMArena Spanish | 1472 | 1234 |
Instruction Following Kimi K3 leads
Kimi K3: 77.7 (#14), Phi-4: 60.4 (#251)
| Benchmark | Kimi K3 | Phi-4 |
|---|---|---|
| LMArena Instruction Following | 1483 | 1201 |
| LiveBench Instruction Following | — | 58.4% |
Long Context Kimi K3 leads
Kimi K3: 45.8 (#29), Phi-4: 36.9 (#226)
| Benchmark | Kimi K3 | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1494 | 1217 |
Writing & Preference Kimi K3 leads
Kimi K3: 76.6 (#4), Phi-4: 40.5 (#244)
| Benchmark | Kimi K3 | Phi-4 |
|---|---|---|
| LMArena Text | 1476 | 1217 |
| LMArena Creative Writing | 1454 | 1182 |
| LMArena Multi-Turn | 1488 | 1206 |
| Short-Story Creative Writing | — | 62.6% |
| EQ-Bench Creative Writing | 2082 | — |
| EQ-Bench 4 | 1339 | — |
| LiveBench Language | — | 25.6% |
Frequently asked questions
Is Kimi K3 better than Phi-4?
Kimi K3 is the stronger model overall, scoring 59.5 to 31.2 on the Noometry Index. Phi-4 costs 69× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Which is cheaper, Kimi K3 or Phi-4?
Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Kimi K3 lists at $3 and $15.
Is Kimi K3 or Phi-4 better for coding?
Kimi K3 scores higher on coding benchmarks: 61.0 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
Kimi K3 does, with 1.05M tokens against 128K.
How many benchmarks do Kimi K3 and Phi-4 share?
21 benchmarks have published results for both models. Kimi K3 has 53 scored results on Noometry and Phi-4 has 37.