Model comparison
GPT-5.6 Terra vs Phi-4 Mini
GPT-5.6 Terra is the stronger model overall, scoring 59.2 to 30.9 on the Noometry Index. Phi-4 Mini costs 34× less per token, which makes it the better buy when GPT-5.6 Terra's lead doesn't matter for your workload.
Last verified . 2 shared benchmarks.
MicrosoftSummary
- They share 2 benchmarks with published results for both. GPT-5.6 Terra scores higher in 3 categories and Phi-4 Mini in 0 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.6 Terra leads 60.7 to 22.4.
- The biggest single-benchmark swing is SciCode: 55% for GPT-5.6 Terra and 10.8% for Phi-4 Mini.
- Phi-4 Mini is cheaper at $0.075 / $0.30 per million input/output tokens, against $2 / $12 for GPT-5.6 Terra.
- GPT-5.6 Terra accepts more context: 1.05M tokens versus 128K.
- Phi-4 Mini has downloadable open weights; the other is API-only.
Side by side
| GPT-5.6 Terra | Phi-4 Mini | |
|---|---|---|
| Provider | OpenAI | Microsoft |
| Noometry Index | 59.2 | 30.9 |
| Released | 2026-07-09 | 2024-12-11 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 128K |
| Max output | 128K | 4K |
| Input $ / M tokens | $2 | $0.075 |
| Output $ / M tokens | $12 | $0.30 |
| Results tracked | 52 | 3 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Terra leads
GPT-5.6 Terra: 57.7 (#19), Phi-4 Mini: 28.1 (#317)
| Benchmark | GPT-5.6 Terra | Phi-4 Mini |
|---|---|---|
| SciCode | 55% | 10.8% |
| DeepSWE | 69.6% | — |
| FrontierCode | 41.3% | — |
| CursorBench | 41.3% | — |
| LMArena WebDev | 1522 | — |
| WeirdML | 78.3% | — |
| LMArena Coding | 1484 | — |
| ALE-Bench | 1,951 | — |
Agentic & Tool Use Not comparable
GPT-5.6 Terra: 40.1 (#25), Phi-4 Mini: —
| Benchmark | GPT-5.6 Terra | Phi-4 Mini |
|---|---|---|
| APEX-Agents | 58.2% | — |
| BALROG | 53.2% | — |
| GDP.pdf | 24.7% | — |
| Vending-Bench 2 | 7,343 | — |
Reasoning GPT-5.6 Terra leads
GPT-5.6 Terra: 60.7 (#21), Phi-4 Mini: 22.4 (#195)
| Benchmark | GPT-5.6 Terra | Phi-4 Mini |
|---|---|---|
| CritPt | 30% | 0% |
| ARC-AGI-2 | 83.9% | — |
| SimpleBench | 48.9% | — |
| Kagi LLM Benchmark | 51.3% | — |
| NYT Connections (extended) | 78.4% | — |
| ARC-AGI-1 | 96.5% | — |
| Chess Puzzles | 54% | — |
| LMArena Hard Prompts | 1468 | — |
| Mystery Game Puzzles | 35% | — |
| DTBench | 93.3% | — |
| LMCA | 55% | — |
| Surface Evolver Bench | 83.8% | — |
| Epoch Capabilities Index | 159.62 | — |
Math Not comparable
GPT-5.6 Terra: 81.6 (#12), Phi-4 Mini: —
| Benchmark | GPT-5.6 Terra | Phi-4 Mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 86% | — |
| FrontierMath Tier 4 | 70.7% | — |
| OTIS Mock AIME 2024-2025 | 99.7% | — |
| ProofBench | 74% | — |
| LMArena Math | 1466 | — |
Knowledge GPT-5.6 Terra leads
GPT-5.6 Terra: 61.2 (#30), Phi-4 Mini: 25.3 (#262)
| Benchmark | GPT-5.6 Terra | Phi-4 Mini |
|---|---|---|
| GPQA Diamond | 93.3% | — |
| SimpleQA Verified | 43.2% | — |
| Vectara Hallucination Rate | — | 23.5% |
| LMArena Expert | 1492 | — |
Multimodal Not comparable
GPT-5.6 Terra: 47.3 (#11), Phi-4 Mini: —
| Benchmark | GPT-5.6 Terra | Phi-4 Mini |
|---|---|---|
| LMArena Vision | 1271 | — |
| Blueprint-Bench 2 | 30.8% | — |
| Furniture Assembly | 54.2% | — |
| LMArena Document | 1472 | — |
Multilingual Not comparable
GPT-5.6 Terra: 54.4 (#44), Phi-4 Mini: —
| Benchmark | GPT-5.6 Terra | Phi-4 Mini |
|---|---|---|
| LMArena Non-English | 1439 | — |
| LMArena Chinese | 1513 | — |
| LMArena French | 1471 | — |
| LMArena German | 1460 | — |
| LMArena Japanese | 1457 | — |
| LMArena Korean | 1425 | — |
| LMArena Russian | 1450 | — |
| LMArena Spanish | 1448 | — |
Instruction Following Not comparable
GPT-5.6 Terra: 76.4 (#40), Phi-4 Mini: —
| Benchmark | GPT-5.6 Terra | Phi-4 Mini |
|---|---|---|
| LMArena Instruction Following | 1454 | — |
Long Context Not comparable
GPT-5.6 Terra: 44.4 (#68), Phi-4 Mini: —
| Benchmark | GPT-5.6 Terra | Phi-4 Mini |
|---|---|---|
| LMArena Longer Query | 1451 | — |
Writing & Preference Not comparable
GPT-5.6 Terra: 70.2 (#23), Phi-4 Mini: —
| Benchmark | GPT-5.6 Terra | Phi-4 Mini |
|---|---|---|
| LMArena Text | 1447 | — |
| LMArena Creative Writing | 1410 | — |
| EQ-Bench Creative Writing | 1855 | — |
| EQ-Bench 4 | 1234 | — |
| LMArena Multi-Turn | 1449 | — |
Frequently asked questions
Is GPT-5.6 Terra better than Phi-4 Mini?
GPT-5.6 Terra is the stronger model overall, scoring 59.2 to 30.9 on the Noometry Index. Phi-4 Mini costs 34× less per token, which makes it the better buy when GPT-5.6 Terra's lead doesn't matter for your workload.
Which is cheaper, GPT-5.6 Terra or Phi-4 Mini?
Phi-4 Mini is cheaper. It lists at $0.075 per million input tokens and $0.30 per million output tokens; GPT-5.6 Terra lists at $2 and $12.
Is GPT-5.6 Terra or Phi-4 Mini better for coding?
GPT-5.6 Terra scores higher on coding benchmarks: 57.7 versus 28.1 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Terra does, with 1.05M tokens against 128K.
How many benchmarks do GPT-5.6 Terra and Phi-4 Mini share?
2 benchmarks have published results for both models. GPT-5.6 Terra has 52 scored results on Noometry and Phi-4 Mini has 3.