DeepSeek V3 vs GLM 4.5
Updated October 21, 2026. 7 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.
Decision reading
DeepSeek V3 scores 70.2 and GLM 4.5 64.2 on the Vibe Coding Index, but no coding benchmark measured both, so the gap is directional, not a ranking.
Pick DeepSeek V3 when output tokens cost 2.1x less, or when the context window is larger, 164K against 131K tokens; pick GLM 4.5 when what matters is Design Arena: websites or Design Arena: UI components, where it leads.
- Shared results
- 7
- DeepSeek V3 only
- 6
- GLM 4.5 only
- 53
- Like-for-like categories
- 1 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
No call
70.2 against 64.2, from different benchmarks with none in common.
Directional only
Intelligence
AA Intelligence Index
No call
Only GLM 4.5 has a AA Intelligence Index result, 12.8.
Directional only
Reasoning
GPQA Diamond
No call
Only GLM 4.5 has a GPQA Diamond result, 72.2%.
Directional only
Math
AIME
No call
Only GLM 4.5 has a AIME result, 86.7%.
Directional only
Knowledge
MMLU Pro
No call
Only GLM 4.5 has a MMLU Pro result, 81.2%.
Directional only
Web design
Design Arena: websites
GLM 4.5
GLM 4.5 leads, 1170 against 1120.
Same board, both measured
Long documents
Listed context window
DeepSeek V3
DeepSeek V3 takes 164K tokens against 131K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
DeepSeek V3
DeepSeek V3 $0.0008, GLM 4.5 $0.0017 per turn.
List prices
At a glance
- Coding index
- DeepSeek V3 70.2 from 1 benchmark, GLM 4.5 64.2 from 1 benchmark.
- Widest category
- Web design, 9.3 points apart on average over 7 shared benchmarks.
- Widest single gap
- Design Arena: SVG: DeepSeek V3 976 against GLM 4.5 1096, 15.0 points apart on a 0 to 100 scale.
- Output price
- DeepSeek V3 $1.0287 against GLM 4.5 $2.20 per million output tokens, 2.1x apart.
- Context window
- DeepSeek V3 takes 164K tokens, GLM 4.5 131K.
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 | DeepSeek V3 | Bars | GLM 4.5 | Ties |
|---|---|---|---|---|
| Web design | 0 | 7 | ||
| All shared boards | 0 | 7 |
CodingDirectional onlyFour of these boards feed the Vibe Coding Index10 boards, none shared
| Benchmark | DeepSeek V3 | GLM 4.5 | Leader |
|---|---|---|---|
| Aider polyglot | 70.2% | Not measured | Only DeepSeek V3 measured |
| SWE-bench Verified | Not measured | 64.2% | Only GLM 4.5 measured |
| SWE-bench Lite | 30.7% | Not measured | Only DeepSeek V3 measured |
| LiveCodeBench (Vals AI) | Not measured | 67.5% | Only GLM 4.5 measured |
| LiveCodeBench (Artificial Analysis run) | Not measured | 73.8% | Only GLM 4.5 measured |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 22.0% | Only GLM 4.5 measured |
| WeirdML | Not measured | 40.6% | Only GLM 4.5 measured |
| ALE-Bench | Not measured | 344.8 | Only GLM 4.5 measured |
| LMArena Text: coding | Not measured | 1434 | Only GLM 4.5 measured |
| LMArena Text: software and IT | Not measured | 1440 | Only GLM 4.5 measured |
| Neither model has a result on the other 28 boards here | |||
IntelligenceDirectional onlyPublic boards, shown as published, kept out of the index1 board, none shared
| Benchmark | DeepSeek V3 | GLM 4.5 | Leader |
|---|---|---|---|
| AA Intelligence Index | Not measured | 12.8 | Only GLM 4.5 measured |
| Neither model has a result on the other 1 board here | |||
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index1 board, none shared
| Benchmark | DeepSeek V3 | GLM 4.5 | Leader |
|---|---|---|---|
| τ²-Bench Telecom | Not measured | 43.0% | Only GLM 4.5 measured |
| Neither model has a result on the other 38 boards here | |||
ReasoningDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
| Benchmark | DeepSeek V3 | GLM 4.5 | Leader |
|---|---|---|---|
| GPQA Diamond | Not measured | 72.2% | Only GLM 4.5 measured |
| GPQA Diamond (Artificial Analysis run) | Not measured | 78.2% | Only GLM 4.5 measured |
| Humanity's Last Exam (Artificial Analysis run) | Not measured | 13.0% | Only GLM 4.5 measured |
| GPQA Diamond (Epoch AI run) | 71.2% | Not measured | Only DeepSeek V3 measured |
| LMArena Text: hard prompts | Not measured | 1429 | Only GLM 4.5 measured |
| LMArena Text: expert | Not measured | 1433 | Only GLM 4.5 measured |
| LMArena Text: hard prompts in English | Not measured | 1429 | Only GLM 4.5 measured |
| Neither model has a result on the other 14 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
| Benchmark | DeepSeek V3 | GLM 4.5 | Leader |
|---|---|---|---|
| AIME | Not measured | 86.7% | Only GLM 4.5 measured |
| MATH 500 (Vals AI) | Not measured | 94.0% | Only GLM 4.5 measured |
| MGSM (Vals AI) | Not measured | 90.8% | Only GLM 4.5 measured |
| AA Math Index | Not measured | 73.7 | Only GLM 4.5 measured |
| AIME 2025 (Artificial Analysis run) | Not measured | 73.7% | Only GLM 4.5 measured |
| OTIS Mock AIME | 48.9% | Not measured | Only DeepSeek V3 measured |
| LMArena Text: math | Not measured | 1427 | Only GLM 4.5 measured |
| LMArena Text: mathematical work | Not measured | 1430 | Only GLM 4.5 measured |
| Neither model has a result on the other 5 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index10 boards, none shared
| Benchmark | DeepSeek V3 | GLM 4.5 | Leader |
|---|---|---|---|
| MMLU Pro | Not measured | 81.2% | Only GLM 4.5 measured |
| LegalBench (Vals AI) | Not measured | 75.6% | Only GLM 4.5 measured |
| CorpFin (Vals AI) | Not measured | 61.0% | Only GLM 4.5 measured |
| TaxEval (Vals AI) | Not measured | 72.4% | Only GLM 4.5 measured |
| MedQA (Vals AI) | Not measured | 90.0% | Only GLM 4.5 measured |
| MMLU Pro (Artificial Analysis run) | Not measured | 83.5% | Only GLM 4.5 measured |
| LMArena Text: science | Not measured | 1444 | Only GLM 4.5 measured |
| LMArena Text: medicine and healthcare | Not measured | 1455 | Only GLM 4.5 measured |
| LMArena Text: legal and government | Not measured | 1438 | Only GLM 4.5 measured |
| LMArena Text: business and finance | Not measured | 1420 | Only GLM 4.5 measured |
| Neither model has a result on the other 7 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index07
| Benchmark | DeepSeek V3 | GLM 4.5 | Leader |
|---|---|---|---|
| Design Arena: websites | 1120 | 1170 | GLM 4.5 |
| Design Arena: UI components | 1093 | 1144 | GLM 4.5 |
| Design Arena: game dev | 1054 | 1145 | GLM 4.5 |
| Design Arena: data viz | 1089 | 1159 | GLM 4.5 |
| Design Arena: 3D | 1087 | 1170 | GLM 4.5 |
| Design Arena: all code categories | 1108 | 1165 | GLM 4.5 |
| Design Arena: SVG | 976 | 1096 | GLM 4.5 |
| Neither model has a result on the other 2 boards here | |||
GeneralDirectional onlyPublic boards, shown as published, kept out of the index22 boards, none shared
| Benchmark | DeepSeek V3 | GLM 4.5 | Leader |
|---|---|---|---|
| IFBench | Not measured | 44.1% | Only GLM 4.5 measured |
| Long context reasoning (AA-LCR) | Not measured | 52.7% | Only GLM 4.5 measured |
| DTBench | 62.7% | Not measured | Only DeepSeek V3 measured |
| LMCA | 15.2% | Not measured | Only DeepSeek V3 measured |
| Creative writing (Lech Mazur) | Not measured | 73.4% | Only GLM 4.5 measured |
| LMArena Text | Not measured | 1430 | Only GLM 4.5 measured |
| LMArena Text: instruction following | Not measured | 1404 | Only GLM 4.5 measured |
| LMArena Text: creative writing | Not measured | 1395 | Only GLM 4.5 measured |
| LMArena Text: English | Not measured | 1437 | Only GLM 4.5 measured |
| LMArena Text: non-English | Not measured | 1417 | Only GLM 4.5 measured |
| LMArena Text: Chinese | Not measured | 1465 | Only GLM 4.5 measured |
| LMArena Text: German | Not measured | 1407 | Only GLM 4.5 measured |
| LMArena Text: Japanese | Not measured | 1415 | Only GLM 4.5 measured |
| LMArena Text: Korean | Not measured | 1380 | Only GLM 4.5 measured |
| LMArena Text: French | Not measured | 1418 | Only GLM 4.5 measured |
| LMArena Text: Spanish | Not measured | 1454 | Only GLM 4.5 measured |
| LMArena Text: Russian | Not measured | 1413 | Only GLM 4.5 measured |
| LMArena Text: Polish | Not measured | 1422 | Only GLM 4.5 measured |
| LMArena Text: multi-turn | Not measured | 1415 | Only GLM 4.5 measured |
| LMArena Text: longer queries | Not measured | 1412 | Only GLM 4.5 measured |
| LMArena Text: writing and language | Not measured | 1403 | Only GLM 4.5 measured |
| LMArena Text: entertainment and media | Not measured | 1404 | Only GLM 4.5 measured |
| Neither model has a result on the other 24 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | DeepSeek V3 | GLM 4.5 | 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
DeepSeek V3 costs less
Repository review
50K input + 3K output tokens
DeepSeek V3 costs less
Agent loop
200K input + 10K output tokens
DeepSeek V3 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 | DeepSeek V3 | Z.ai GLM 4.5 |
|---|---|---|
| Maker | DeepSeek | Z.ai |
| Weights | Open weights | Open weights |
| Released | December 26, 2024 | July 25, 2025 |
| Context window | 164K | 131K |
| Input $/Mtok | $0.2574 | $0.60 |
| Output $/Mtok | $1.0287 | $2.20 |
| Coding index | 70.2 | 64.2 |
| Independent measurements behind it | 1 | 1 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is DeepSeek V3 better than GLM 4.5 for coding?
DeepSeek V3 scores 70.2 and GLM 4.5 64.2 on the Vibe Coding Index, but no coding benchmark measured both, so the gap is directional, not a ranking. Pick DeepSeek V3 when output tokens cost 2.1x less, or when the context window is larger, 164K against 131K tokens; pick GLM 4.5 when what matters is Design Arena: websites or Design Arena: UI components, where it leads.
Which is cheaper, DeepSeek V3 or GLM 4.5?
DeepSeek V3 costs $0.2574 per million input tokens and $1.0287 per million output tokens; GLM 4.5 costs $0.60 and $2.20. Cost per solved task is known for neither of them.
Which has the bigger context window, DeepSeek V3 or GLM 4.5?
DeepSeek V3, with 164K tokens against 131K.
How many benchmarks is the DeepSeek V3 vs GLM 4.5 comparison based on?
7 benchmarks have a result for both models as of October 21, 2026. DeepSeek V3's coding index rests on 1 independent benchmark and GLM 4.5's on 1. Where only one model has a result, the page says so instead of naming a winner.