GPT-4 Turbo vs Phi 4
Updated October 21, 2026. 33 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.
Decision reading
Neither model has a coding score yet, so the two cannot be ranked on coding.
GPT-4 Turbo has the edge where the context window is larger, 128K against 16K tokens, and where what matters is Epoch Capabilities Index or Humanity's Last Exam (Artificial Analysis run), where it leads.
- Shared results
- 33
- GPT-4 Turbo only
- 9
- Phi 4 only
- 12
- Like-for-like categories
- 6 of 9
Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
No call
Neither has a coding score yet.
One side not measured
Intelligence
AA Intelligence Index
GPT-4 Turbo
GPT-4 Turbo leads, 7.0 against 5.9.
Same board, both measured
Long documents
Listed context window
GPT-4 Turbo
GPT-4 Turbo takes 128K tokens against 16K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Phi 4
GPT-4 Turbo $0.025, Phi 4 $0.0001 per turn.
List prices
At a glance
- Coding index
- GPT-4 Turbo not benchmarked, Phi 4 not benchmarked.
- Widest category
- General, 8.8 points apart on average over 16 shared benchmarks.
- Widest single gap
- LMArena Text: entertainment and media: GPT-4 Turbo 1258 against Phi 4 1162, 16.0 points apart on a 0 to 100 scale.
- Output price
- GPT-4 Turbo $30.00 against Phi 4 $0.14 per million output tokens, 214.3x apart.
- Context window
- GPT-4 Turbo takes 128K tokens, Phi 4 16K.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | GPT-4 Turbo | Bars | Phi 4 | Ties |
|---|---|---|---|---|
| Coding | 3 | 0 | ||
| Intelligence | 1 | 1 | ||
| Reasoning | 3 | 1 | ||
| Math | 2 | 0 | ||
| Knowledge | 4 | 2 | ||
| General | 16 | 0 | ||
| All shared boards | 29 | 4 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index30
| Benchmark | GPT-4 Turbo | Phi 4 | Leader |
|---|---|---|---|
| AA Coding Index | 21.5 | Not measured | Only GPT-4 Turbo measured |
| LiveCodeBench (Artificial Analysis run) | 29.1% | 23.1% | GPT-4 Turbo |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 3.8% | Only Phi 4 measured |
| LMArena Text: coding | 1268 | 1231 | GPT-4 Turbo |
| LMArena Text: software and IT | 1267 | 1228 | GPT-4 Turbo |
| Neither model has a result on the other 33 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index11
| Benchmark | GPT-4 Turbo | Phi 4 | Leader |
|---|---|---|---|
| AA Intelligence Index | 7.0 | 5.9 | GPT-4 Turbo |
| Epoch Capabilities Index | 127.3 | 130.5 | Phi 4 |
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
| Benchmark | GPT-4 Turbo | Phi 4 | Leader |
|---|---|---|---|
| τ²-Bench Telecom | Not measured | 0.0% | Only Phi 4 measured |
| BALROG | Not measured | 11.6% | Only Phi 4 measured |
| Neither model has a result on the other 37 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index31
| Benchmark | GPT-4 Turbo | Phi 4 | Leader |
|---|---|---|---|
| GPQA Diamond (Artificial Analysis run) | Not measured | 57.5% | Only Phi 4 measured |
| Humanity's Last Exam (Artificial Analysis run) | 3.1% | 3.8% | Phi 4 |
| GPQA Diamond (Epoch AI run) | Not measured | 56.1% | Only Phi 4 measured |
| LMArena Text: hard prompts | 1251 | 1220 | GPT-4 Turbo |
| LMArena Text: expert | 1223 | 1203 | GPT-4 Turbo |
| LMArena Text: hard prompts in English | 1266 | 1230 | GPT-4 Turbo |
| Neither model has a result on the other 15 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | GPT-4 Turbo | Phi 4 | Leader |
|---|---|---|---|
| AA Math Index | Not measured | 18.0 | Only Phi 4 measured |
| AIME 2025 (Artificial Analysis run) | Not measured | 18.0% | Only Phi 4 measured |
| OTIS Mock AIME | Not measured | 13.8% | Only Phi 4 measured |
| MATH level 5 | Not measured | 64.9% | Only Phi 4 measured |
| LMArena Text: math | 1272 | 1246 | GPT-4 Turbo |
| LMArena Text: mathematical work | 1268 | 1254 | GPT-4 Turbo |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index42
| Benchmark | GPT-4 Turbo | Phi 4 | Leader |
|---|---|---|---|
| LegalBench (Vals AI) | 80.5% | Not measured | Only GPT-4 Turbo measured |
| MedQA (Vals AI) | 82.0% | Not measured | Only GPT-4 Turbo measured |
| MMLU Pro (Artificial Analysis run) | 69.4% | 71.4% | Phi 4 |
| MMLU | 81.3% | 84.8% | Phi 4 |
| LMArena Text: science | 1256 | 1206 | GPT-4 Turbo |
| LMArena Text: medicine and healthcare | 1239 | 1202 | GPT-4 Turbo |
| LMArena Text: legal and government | 1265 | 1241 | GPT-4 Turbo |
| LMArena Text: business and finance | 1222 | 1199 | GPT-4 Turbo |
| Neither model has a result on the other 9 boards here | |||
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GPT-4 Turbo | Phi 4 | Leader |
|---|---|---|---|
| Neither model has a result on 9 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index160
| Benchmark | GPT-4 Turbo | Phi 4 | Leader |
|---|---|---|---|
| IFBench | Not measured | 23.5% | Only Phi 4 measured |
| Long context reasoning (AA-LCR) | Not measured | 0.0% | Only Phi 4 measured |
| DTBench | 61.6% | Not measured | Only GPT-4 Turbo measured |
| LMCA | 9.8% | Not measured | Only GPT-4 Turbo measured |
| Creative writing (Lech Mazur) | Not measured | 62.6% | Only Phi 4 measured |
| LMArena Text | 1272 | 1217 | GPT-4 Turbo |
| LMArena Text: instruction following | 1249 | 1201 | GPT-4 Turbo |
| LMArena Text: creative writing | 1268 | 1181 | GPT-4 Turbo |
| LMArena Vision | 1090 | Not measured | Only GPT-4 Turbo measured |
| LMArena Text: English | 1291 | 1231 | GPT-4 Turbo |
| LMArena Text: non-English | 1245 | 1197 | GPT-4 Turbo |
| LMArena Text: Chinese | 1242 | 1212 | GPT-4 Turbo |
| LMArena Text: German | 1259 | 1222 | GPT-4 Turbo |
| LMArena Text: Japanese | 1194 | 1158 | GPT-4 Turbo |
| LMArena Text: Korean | 1187 | 1151 | GPT-4 Turbo |
| LMArena Text: French | 1276 | 1224 | GPT-4 Turbo |
| LMArena Text: Spanish | 1260 | 1234 | GPT-4 Turbo |
| LMArena Text: Russian | 1259 | 1209 | GPT-4 Turbo |
| LMArena Text: Polish | 1302 | Not measured | Only GPT-4 Turbo measured |
| LMArena Text: multi-turn | 1267 | 1206 | GPT-4 Turbo |
| LMArena Text: longer queries | 1254 | 1217 | GPT-4 Turbo |
| LMArena Text: writing and language | 1279 | 1198 | GPT-4 Turbo |
| LMArena Text: entertainment and media | 1258 | 1162 | GPT-4 Turbo |
| LMArena Vision: English | 1097 | Not measured | Only GPT-4 Turbo measured |
| LMArena Vision: Chinese | 1045 | Not measured | Only GPT-4 Turbo measured |
| Neither model has a result on the other 21 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GPT-4 Turbo | Phi 4 | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
Phi 4 costs less
Repository review
50K input + 3K output tokens
Phi 4 costs less
Agent loop
200K input + 10K output tokens
Phi 4 costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | OpenAI GPT-4 Turbo | Microsoft Phi 4 |
|---|---|---|
| Maker | OpenAI | Microsoft |
| Weights | No public weights listed | Open weights |
| Released | April 9, 2024 | January 10, 2025 |
| Context window | 128K | 16K |
| Input $/Mtok | $10.00 | $0.07 |
| Output $/Mtok | $30.00 | $0.14 |
| Coding index | Not benchmarked | Not benchmarked |
| Independent measurements behind it | 0 | 0 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GPT-4 Turbo better than Phi 4 for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. GPT-4 Turbo has the edge where the context window is larger, 128K against 16K tokens, and where what matters is Epoch Capabilities Index or Humanity's Last Exam (Artificial Analysis run), where it leads.
Which is cheaper, GPT-4 Turbo or Phi 4?
GPT-4 Turbo costs $10.00 per million input tokens and $30.00 per million output tokens; Phi 4 costs $0.07 and $0.14. Cost per solved task is known for neither of them.
Which has the bigger context window, GPT-4 Turbo or Phi 4?
GPT-4 Turbo, with 128K tokens against 16K.
How many benchmarks is the GPT-4 Turbo vs Phi 4 comparison based on?
33 benchmarks have a result for both models as of October 21, 2026. GPT-4 Turbo's coding index rests on 0 independent benchmarks and Phi 4's on 0. Where only one model has a result, the page says so instead of naming a winner.