Model comparison
Grok 4.20 (Non-Reasoning) vs Phi-4
Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.6 to 31.2 on the Noometry Index. Phi-4 costs 18× less per token, which makes it the better buy when Grok 4.20 (Non-Reasoning)'s lead doesn't matter for your workload.
Last verified . 21 shared benchmarks.
MicrosoftSummary
- They share 21 benchmarks with published results for both. Grok 4.20 (Non-Reasoning) scores higher in 9 categories and Phi-4 in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.20 (Non-Reasoning) leads 52.3 to 17.7.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 92.2% for Grok 4.20 (Non-Reasoning) and 13.8% for Phi-4.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $1.25 / $2.50 for Grok 4.20 (Non-Reasoning).
- Grok 4.20 (Non-Reasoning) accepts more context: 1M tokens versus 128K.
- Phi-4 has downloadable open weights; the other is API-only.
Side by side
| Grok 4.20 (Non-Reasoning) | Phi-4 | |
|---|---|---|
| Provider | xAI | Microsoft |
| Noometry Index | 48.6 | 31.2 |
| Released | 2026-02-17 | 2024-12-11 |
| Weights | Proprietary | Open |
| Context window | 1M | 128K |
| Max output | 30K | 4K |
| Input $ / M tokens | $1.25 | $0.07 |
| Output $ / M tokens | $2.50 | $0.14 |
| Results tracked | 46 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 42.1 (#112), Phi-4: 34.4 (#239)
| Benchmark | Grok 4.20 (Non-Reasoning) | Phi-4 |
|---|---|---|
| LMArena Coding | 1459 | 1231 |
| LMArena WebDev | 1375 | — |
| WeirdML | 52.3% | — |
| BigCodeBench Instruct | — | 45.5% |
| LiveBench Coding | — | 30.7% |
| BigCodeBench Complete | — | 55.4% |
| ALE-Bench | 1,150 | — |
Agentic & Tool Use Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 34.4 (#46), Phi-4: 22.8 (#128)
| Benchmark | Grok 4.20 (Non-Reasoning) | Phi-4 |
|---|---|---|
| Terminal-Bench | 57.3% | — |
| Berkeley Function Calling Leaderboard | — | 28.8% |
| τ²-bench Banking | 18% | — |
| BALROG | — | 11.6% |
| LMArena Search | 1189 | — |
| Vending-Bench 2 | 4,663 | — |
Reasoning Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 52.3 (#32), Phi-4: 17.7 (#291)
| Benchmark | Grok 4.20 (Non-Reasoning) | Phi-4 |
|---|---|---|
| Chess Puzzles | 24% | 1% |
| LMArena Hard Prompts | 1451 | 1220 |
| Epoch Capabilities Index | 151.98 | 130.42 |
| ARC-AGI-2 | 65.1% | — |
| Kagi LLM Benchmark | 75% | — |
| NYT Connections (extended) | 85.4% | — |
| ARC-AGI-1 | 89.5% | — |
| Thematic Generalization | 63.8% | — |
| LiveBench Reasoning | — | 47.8% |
| DTBench | 90.1% | — |
| LiveBench Data Analysis | — | 45.2% |
| LMCA | 38.7% | — |
| ForecastBench | 61.4 | — |
| LiveBench | — | 41.6% |
Math Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 48.2 (#65), Phi-4: 20.8 (#285)
| Benchmark | Grok 4.20 (Non-Reasoning) | Phi-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 92.2% | 13.8% |
| LMArena Math | 1455 | 1246 |
| FrontierMath (Tiers 1-3) | 44.9% | — |
| FrontierMath Tier 4 | 17.1% | — |
| ProofBench | 14% | — |
| LiveBench Math | — | 42% |
| MATH Level 5 | — | 64.9% |
Knowledge Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 52.8 (#60), Phi-4: 32.6 (#209)
| Benchmark | Grok 4.20 (Non-Reasoning) | Phi-4 |
|---|---|---|
| GPQA Diamond | 89.3% | 56.1% |
| LMArena Expert | 1439 | 1203 |
| SimpleQA Verified | 30.2% | — |
| Confabulations | — | 29.4% |
| Vectara Hallucination Rate | — | 3.7% |
| MMLU | — | 84.8% |
Multimodal Not comparable
Grok 4.20 (Non-Reasoning): 33.3 (#98), Phi-4: —
| Benchmark | Grok 4.20 (Non-Reasoning) | Phi-4 |
|---|---|---|
| LMArena Vision | 1263 | — |
| Blueprint-Bench 2 | 0% | — |
| LMArena Document | 1416 | — |
Multilingual Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 54.5 (#40), Phi-4: 37.2 (#237)
| Benchmark | Grok 4.20 (Non-Reasoning) | Phi-4 |
|---|---|---|
| LMArena Non-English | 1441 | 1197 |
| LMArena Chinese | 1481 | 1212 |
| LMArena French | 1476 | 1224 |
| LMArena German | 1465 | 1222 |
| LMArena Japanese | 1449 | 1158 |
| LMArena Korean | 1417 | 1151 |
| LMArena Russian | 1458 | 1209 |
| LMArena Spanish | 1443 | 1234 |
Instruction Following Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 74.8 (#83), Phi-4: 60.4 (#251)
| Benchmark | Grok 4.20 (Non-Reasoning) | Phi-4 |
|---|---|---|
| LMArena Instruction Following | 1420 | 1201 |
| LiveBench Instruction Following | — | 58.4% |
Long Context Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 45.5 (#34), Phi-4: 36.9 (#226)
| Benchmark | Grok 4.20 (Non-Reasoning) | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1437 | 1217 |
| CL-bench | 22.2% | — |
| CL-bench Life | 11.9% | — |
Writing & Preference Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 65.7 (#44), Phi-4: 40.5 (#244)
| Benchmark | Grok 4.20 (Non-Reasoning) | Phi-4 |
|---|---|---|
| LMArena Text | 1451 | 1217 |
| LMArena Creative Writing | 1438 | 1182 |
| LMArena Multi-Turn | 1456 | 1206 |
| Short-Story Creative Writing | — | 62.6% |
| EQ-Bench Creative Writing | 1574 | — |
| LiveBench Language | — | 25.6% |
Frequently asked questions
Is Grok 4.20 (Non-Reasoning) better than Phi-4?
Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.6 to 31.2 on the Noometry Index. Phi-4 costs 18× less per token, which makes it the better buy when Grok 4.20 (Non-Reasoning)'s lead doesn't matter for your workload.
Which is cheaper, Grok 4.20 (Non-Reasoning) or Phi-4?
Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Grok 4.20 (Non-Reasoning) lists at $1.25 and $2.50.
Is Grok 4.20 (Non-Reasoning) or Phi-4 better for coding?
Grok 4.20 (Non-Reasoning) scores higher on coding benchmarks: 42.1 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
Grok 4.20 (Non-Reasoning) does, with 1M tokens against 128K.
How many benchmarks do Grok 4.20 (Non-Reasoning) and Phi-4 share?
21 benchmarks have published results for both models. Grok 4.20 (Non-Reasoning) has 46 scored results on Noometry and Phi-4 has 37.