GLM 4.7 vs GPT-4
Updated October 21, 2026. 34 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.
Decision reading
Pick GLM 4.7 for coding: it scores 33.6 on the Vibe Coding Index against 22.4 for GPT-4.
GPT-4 is neither cheaper nor larger in context, and leads no benchmark both models were measured on.
- Shared results
- 34
- GLM 4.7 only
- 46
- GPT-4 only
- 11
- 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, four public coding benchmarks
GLM 4.7
GLM 4.7 scores 33.6 against 22.4.
One benchmark behind each score
Intelligence
AA Intelligence Index
GLM 4.7
GLM 4.7 leads, 22.2 against 6.7.
Same board, both measured
Reasoning
GPQA Diamond
No call
Only GLM 4.7 has a GPQA Diamond result, 80.0%.
Directional only
Math
AIME
No call
Only GLM 4.7 has a AIME result, 93.3%.
Directional only
Knowledge
MMLU Pro
No call
Only GLM 4.7 has a MMLU Pro result, 82.7%.
Directional only
Web design
Design Arena: websites
No call
Only GLM 4.7 has a Design Arena: websites result, 1226.
Directional only
Long documents
Listed context window
GLM 4.7
GLM 4.7 takes 205K tokens against 8K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
GLM 4.7
GLM 4.7 $0.0015, GPT-4 $0.060 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only GLM 4.7 has the number, $0.29.
Directional only
At a glance
- Coding index
- GLM 4.7 33.6 from 1 benchmark, GPT-4 22.4 from 1 benchmark.
- Widest category
- Reasoning, 44.5 points apart on average over 4 shared benchmarks.
- Widest single gap
- LMArena Text: medicine and healthcare: GLM 4.7 1469 against GPT-4 1145, 53.9 points apart on a 0 to 100 scale.
- Output price
- GLM 4.7 $1.98 against GPT-4 $60.00 per million output tokens, 30.3x apart.
- Context window
- GLM 4.7 takes 205K tokens, GPT-4 8K.
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 | GLM 4.7 | Bars | GPT-4 | Ties |
|---|---|---|---|---|
| Coding | 3 | 0 | ||
| Intelligence | 2 | 0 | ||
| Reasoning | 4 | 0 | ||
| Math | 2 | 0 | ||
| Knowledge | 5 | 0 | ||
| General | 18 | 0 | ||
| All shared boards | 34 | 0 |
CodingLike-for-likeFour of these boards feed the Vibe Coding Index30
| Benchmark | GLM 4.7 | GPT-4 | Leader |
|---|---|---|---|
| SWE-bench Verified | Not measured | 22.4% | Only GPT-4 measured |
| SWE-bench Lite | Not measured | 39.7% | Only GPT-4 measured |
| SWE-bench (full test set) | Not measured | 12.5% | Only GPT-4 measured |
| LMArena WebDev | 1435 | Not measured | Only GLM 4.7 measured |
| AA Coding Index | 45.3 | 13.1 | GLM 4.7 |
| LiveCodeBench (Vals AI) | 82.2% | Not measured | Only GLM 4.7 measured |
| SWE-bench Verified (Vals AI) | 69.4% | Not measured | Only GLM 4.7 measured |
| LiveCodeBench (Artificial Analysis run) | 89.4% | Not measured | Only GLM 4.7 measured |
| Terminal-Bench Hard (Artificial Analysis run) | 31.8% | Not measured | Only GLM 4.7 measured |
| Terminal-Bench | 33.4% | Not measured | Only GLM 4.7 measured |
| ALE-Bench | 399.5 | Not measured | Only GLM 4.7 measured |
| LMArena Text: coding | 1454 | 1209 | GLM 4.7 |
| LMArena Text: software and IT | 1453 | 1195 | GLM 4.7 |
| LMArena WebDev: React | 1420 | Not measured | Only GLM 4.7 measured |
| LMArena WebDev: HTML | 1447 | Not measured | Only GLM 4.7 measured |
| Neither model has a result on the other 23 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | GLM 4.7 | GPT-4 | Leader |
|---|---|---|---|
| AA Intelligence Index | 22.2 | 6.7 | GLM 4.7 |
| Epoch Capabilities Index | 143.6 | 125.7 | GLM 4.7 |
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
| Benchmark | GLM 4.7 | GPT-4 | Leader |
|---|---|---|---|
| Design Arena: agents building full-stack apps | 1050 | Not measured | Only GLM 4.7 measured |
| Design Arena: agents building mobile apps | 1105 | Not measured | Only GLM 4.7 measured |
| Design Arena: agents building Android apps | 1118 | Not measured | Only GLM 4.7 measured |
| Design Arena: agents building Godot games | 1058 | Not measured | Only GLM 4.7 measured |
| τ²-Bench Telecom | 95.9% | Not measured | Only GLM 4.7 measured |
| Vending-Bench 2 | $2,377 | Not measured | Only GLM 4.7 measured |
| Neither model has a result on the other 33 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index40
| Benchmark | GLM 4.7 | GPT-4 | Leader |
|---|---|---|---|
| GPQA Diamond | 80.0% | Not measured | Only GLM 4.7 measured |
| GPQA Diamond (Artificial Analysis run) | 85.9% | 34.9% | GLM 4.7 |
| Humanity's Last Exam (Artificial Analysis run) | 27.4% | Not measured | Only GLM 4.7 measured |
| GPQA Diamond (Epoch AI run) | 83.3% | Not measured | Only GLM 4.7 measured |
| SimpleBench | 47.7% | Not measured | Only GLM 4.7 measured |
| Chess puzzles | 21.0% | Not measured | Only GLM 4.7 measured |
| BIG-Bench Hard | Not measured | 75.1% | Only GPT-4 measured |
| LMArena Text: hard prompts | 1443 | 1200 | GLM 4.7 |
| LMArena Text: expert | 1424 | 1149 | GLM 4.7 |
| LMArena Text: hard prompts in English | 1452 | 1209 | GLM 4.7 |
| Neither model has a result on the other 11 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | GLM 4.7 | GPT-4 | Leader |
|---|---|---|---|
| AIME | 93.3% | Not measured | Only GLM 4.7 measured |
| MGSM (Vals AI) | 88.2% | Not measured | Only GLM 4.7 measured |
| AA Math Index | 95.0 | Not measured | Only GLM 4.7 measured |
| AIME 2025 (Artificial Analysis run) | 95.0% | Not measured | Only GLM 4.7 measured |
| OTIS Mock AIME | 83.3% | Not measured | Only GLM 4.7 measured |
| ProofBench | 6.0% | Not measured | Only GLM 4.7 measured |
| GSM8K | Not measured | 92.0% | Only GPT-4 measured |
| LMArena Text: math | 1424 | 1230 | GLM 4.7 |
| LMArena Text: mathematical work | 1411 | 1228 | GLM 4.7 |
| Neither model has a result on the other 4 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index50
| Benchmark | GLM 4.7 | GPT-4 | Leader |
|---|---|---|---|
| MMLU Pro | 82.7% | Not measured | Only GLM 4.7 measured |
| LegalBench (Vals AI) | 83.4% | Not measured | Only GLM 4.7 measured |
| CaseLaw (Vals AI) | 54.9% | Not measured | Only GLM 4.7 measured |
| CorpFin (Vals AI) | 46.4% | Not measured | Only GLM 4.7 measured |
| TaxEval (Vals AI) | 68.8% | Not measured | Only GLM 4.7 measured |
| MedQA (Vals AI) | 93.7% | Not measured | Only GLM 4.7 measured |
| MedCode (Vals AI) | 32.8% | Not measured | Only GLM 4.7 measured |
| MMLU Pro (Artificial Analysis run) | 85.6% | 56.2% | GLM 4.7 |
| SimpleQA Verified | 32.2% | Not measured | Only GLM 4.7 measured |
| MMLU | Not measured | 86.4% | Only GPT-4 measured |
| AI2 Reasoning Challenge | Not measured | 96.3% | Only GPT-4 measured |
| LMArena Text: science | 1457 | 1177 | GLM 4.7 |
| LMArena Text: medicine and healthcare | 1469 | 1145 | GLM 4.7 |
| LMArena Text: legal and government | 1444 | 1191 | GLM 4.7 |
| LMArena Text: business and finance | 1429 | 1145 | GLM 4.7 |
| Neither model has a result on the other 2 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
| Benchmark | GLM 4.7 | GPT-4 | Leader |
|---|---|---|---|
| Design Arena: websites | 1226 | Not measured | Only GLM 4.7 measured |
| Design Arena: UI components | 1199 | Not measured | Only GLM 4.7 measured |
| Design Arena: game dev | 1189 | Not measured | Only GLM 4.7 measured |
| Design Arena: data viz | 1195 | Not measured | Only GLM 4.7 measured |
| Design Arena: 3D | 1191 | Not measured | Only GLM 4.7 measured |
| Design Arena: all code categories | 1217 | Not measured | Only GLM 4.7 measured |
| Design Arena: SVG | 1142 | Not measured | Only GLM 4.7 measured |
| Design Arena: ASCII art | 1165 | Not measured | Only GLM 4.7 measured |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index180
| Benchmark | GLM 4.7 | GPT-4 | Leader |
|---|---|---|---|
| MedScribe (Vals AI) | 68.6% | Not measured | Only GLM 4.7 measured |
| IFBench | 67.9% | 33.2% | GLM 4.7 |
| Long context reasoning (AA-LCR) | 71.0% | Not measured | Only GLM 4.7 measured |
| DTBench | Not measured | 62.7% | Only GPT-4 measured |
| LMCA | Not measured | 17.1% | Only GPT-4 measured |
| CL-bench | 15.9% | Not measured | Only GLM 4.7 measured |
| HellaSwag | Not measured | 95.3% | Only GPT-4 measured |
| WinoGrande | Not measured | 87.5% | Only GPT-4 measured |
| CL-bench Life | 10.9% | Not measured | Only GLM 4.7 measured |
| LMArena Text | 1435 | 1206 | GLM 4.7 |
| LMArena Text: instruction following | 1412 | 1201 | GLM 4.7 |
| LMArena Text: creative writing | 1401 | 1191 | GLM 4.7 |
| LMArena Text: English | 1453 | 1220 | GLM 4.7 |
| LMArena Text: non-English | 1417 | 1194 | GLM 4.7 |
| LMArena Text: Chinese | 1495 | 1184 | GLM 4.7 |
| LMArena Text: German | 1426 | 1198 | GLM 4.7 |
| LMArena Text: Japanese | 1439 | 1137 | GLM 4.7 |
| LMArena Text: Korean | 1399 | 1087 | GLM 4.7 |
| LMArena Text: French | 1432 | 1220 | GLM 4.7 |
| LMArena Text: Spanish | 1434 | 1197 | GLM 4.7 |
| LMArena Text: Russian | 1423 | 1195 | GLM 4.7 |
| LMArena Text: Polish | 1435 | 1265 | GLM 4.7 |
| LMArena Text: multi-turn | 1446 | 1206 | GLM 4.7 |
| LMArena Text: longer queries | 1432 | 1190 | GLM 4.7 |
| LMArena Text: writing and language | 1411 | 1219 | GLM 4.7 |
| LMArena Text: entertainment and media | 1400 | 1184 | GLM 4.7 |
| Neither model has a result on the other 20 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GLM 4.7 | GPT-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
GLM 4.7 costs less
Repository review
50K input + 3K output tokens
GLM 4.7 costs less
Agent loop
200K input + 10K output tokens
GLM 4.7 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 | Z.ai GLM 4.7 | OpenAI GPT-4 |
|---|---|---|
| Maker | Z.ai | OpenAI |
| Weights | Open weights | No public weights listed |
| Released | December 22, 2025 | May 28, 2023 |
| Context window | 205K | 8K |
| Input $/Mtok | $0.54 | $30.00 |
| Output $/Mtok | $1.98 | $60.00 |
| Coding index | 33.6 | 22.4 |
| Independent measurements behind it | 1 | 1 |
| Cost per solved task | $0.29 | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GLM 4.7 better than GPT-4 for coding?
Pick GLM 4.7 for coding: it scores 33.6 on the Vibe Coding Index against 22.4 for GPT-4. GPT-4 is neither cheaper nor larger in context, and leads no benchmark both models were measured on.
Which is cheaper, GLM 4.7 or GPT-4?
GLM 4.7 costs $0.54 per million input tokens and $1.98 per million output tokens; GPT-4 costs $30.00 and $60.00. Cost per solved task is known for GLM 4.7 only.
Which has the bigger context window, GLM 4.7 or GPT-4?
GLM 4.7, with 205K tokens against 8K.
How many benchmarks is the GLM 4.7 vs GPT-4 comparison based on?
34 benchmarks have a result for both models as of October 21, 2026. GLM 4.7's coding index rests on 1 independent benchmark and GPT-4's on 1. Where only one model has a result, the page says so instead of naming a winner.