Grok 4.3 vs Qwen2.5 72B Instruct
Grok 4.3 leads Qwen2.5 72B Instruct on the Artificial Analysis Intelligence Index, 24.9 to 7.7, and Qwen2.5 72B Instruct costs 6.3x less per output token ($0.40 vs $2.50 per million). Updated October 21, 2026; 38 benchmarks have a result for both.
Grok 4.3 or Qwen2.5 72B Instruct: which to pick
Only Grok 4.3 has a coding score, 37.4 on the Vibe Coding Index; Qwen2.5 72B Instruct has none, so the two cannot be ranked on coding.
Pick Qwen2.5 72B Instruct for output tokens 6.3x cheaper ($0.40 vs $2.50 per million).
- Shared results
- 38
- Grok 4.3 only
- 66
- Qwen2.5 72B Instruct only
- 15
- Like-for-like categories
- 7 of 9
Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.
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
Only Grok 4.3 has a coding score, so no call is made.
One side not measured
Intelligence
AA Intelligence Index
Grok 4.3
Grok 4.3 leads, 24.9 against 7.7.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Grok 4.3 has a AA Agentic Index result, 15.5.
Directional only
Reasoning
GPQA Diamond
No call
Only Grok 4.3 has a GPQA Diamond result, 91.4%.
Directional only
Knowledge
MMLU Pro
No call
Only Grok 4.3 has a MMLU Pro result, 85.8%.
Directional only
Web design
Design Arena: websites
No call
Only Grok 4.3 has a Design Arena: websites result, 1196.
Directional only
Long documents
Listed context window
Grok 4.3
Grok 4.3 takes 1M tokens against 33K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Qwen2.5 72B Instruct
Grok 4.3 $0.0025, Qwen2.5 72B Instruct $0.0006 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only Grok 4.3 has the number, $0.31.
Directional only
At a glance
- Coding index
- Grok 4.3 37.4 from 3 benchmarks, Qwen2.5 72B Instruct not benchmarked.
- Widest category
- Agentic work, 63.2 points apart on average over 1 shared benchmark.
- Widest single gap
- OTIS Mock AIME: Grok 4.3 93.3% against Qwen2.5 72B Instruct 8.1%, 85.3 points apart on a 0 to 100 scale.
- Output price
- Grok 4.3 $2.50 against Qwen2.5 72B Instruct $0.40 per million output tokens, 6.3x apart.
- Context window
- Grok 4.3 takes 1M tokens, Qwen2.5 72B Instruct 33K.
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 | Grok 4.3 | Bars | Qwen2.5 72B Instruct | Ties |
|---|---|---|---|---|
| Coding | 4 | 0 | ||
| Intelligence | 1 | 0 | ||
| Agentic work | 1 | 0 | ||
| Reasoning | 6 | 0 | ||
| Math | 3 | 0 | ||
| Knowledge | 4 | 0 | ||
| General | 19 | 0 | ||
| All shared boards | 38 | 0 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index40
| Benchmark | Grok 4.3 | Qwen2.5 72B Instruct | Leader |
|---|---|---|---|
| LMArena WebDev | 1358 | Not measured | Only Grok 4.3 measured |
| AA Coding Index | 42.2 | Not measured | Only Grok 4.3 measured |
| LiveCodeBench (Vals AI) | 84.5% | Not measured | Only Grok 4.3 measured |
| SWE-bench Verified (Vals AI) | 71.4% | Not measured | Only Grok 4.3 measured |
| Vibe Code Bench (Vals AI) | 19.4% | Not measured | Only Grok 4.3 measured |
| Code migration (Vals AI) | 6.8% | Not measured | Only Grok 4.3 measured |
| ProgramBench (Vals AI) | 0.0% | Not measured | Only Grok 4.3 measured |
| Terminal-Bench 2.1 (Vals AI) | 42.0% | Not measured | Only Grok 4.3 measured |
| SkillsBench (Vals AI) | 40.6% | Not measured | Only Grok 4.3 measured |
| LiveCodeBench (Artificial Analysis run) | Not measured | 27.6% | Only Qwen2.5 72B Instruct measured |
| SciCode | 48.3% | Not measured | Only Grok 4.3 measured |
| Terminal-Bench Hard (Artificial Analysis run) | 37.9% | 4.5% | Grok 4.3 |
| WeirdML | 49.9% | 16.0% | Grok 4.3 |
| ALE-Bench | 944.2 | Not measured | Only Grok 4.3 measured |
| LMArena Text: coding | 1415 | 1292 | Grok 4.3 |
| LMArena Text: software and IT | 1416 | 1292 | Grok 4.3 |
| LMArena WebDev: React | 1353 | Not measured | Only Grok 4.3 measured |
| LMArena WebDev: HTML | 1367 | Not measured | Only Grok 4.3 measured |
| LMArena WebDev: from an image | 1375 | Not measured | Only Grok 4.3 measured |
| Neither model has a result on the other 20 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Grok 4.3 | Qwen2.5 72B Instruct | Leader |
|---|---|---|---|
| AA Intelligence Index | 24.9 | 7.7 | Grok 4.3 |
| Epoch Capabilities Index | Not measured | 129.0 | Only Qwen2.5 72B Instruct measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Grok 4.3 | Qwen2.5 72B Instruct | Leader |
|---|---|---|---|
| AA Agentic Index | 15.5 | Not measured | Only Grok 4.3 measured |
| Design Arena: agents building web apps | 1126 | Not measured | Only Grok 4.3 measured |
| Design Arena: agents building full-stack apps | 1006 | Not measured | Only Grok 4.3 measured |
| Design Arena: agents building mobile apps | 1059 | Not measured | Only Grok 4.3 measured |
| Design Arena: agents building Android apps | 961 | Not measured | Only Grok 4.3 measured |
| Design Arena: agents building Godot games | 914 | Not measured | Only Grok 4.3 measured |
| Design Arena: agents building games | 1007 | Not measured | Only Grok 4.3 measured |
| Design Arena: HTML slides | 1008 | Not measured | Only Grok 4.3 measured |
| Design Arena: PowerPoint slides | 1013 | Not measured | Only Grok 4.3 measured |
| Finance Agent (Vals AI) | 37.7% | Not measured | Only Grok 4.3 measured |
| Tax Agent Bench (Vals AI) | 6.2% | Not measured | Only Grok 4.3 measured |
| Legal Research Bench (Vals AI) | 15.4% | Not measured | Only Grok 4.3 measured |
| Excel Modeling Benchmark (Vals AI) | 18.4% | Not measured | Only Grok 4.3 measured |
| τ²-Bench Telecom | 97.7% | 34.5% | Grok 4.3 |
| METR time horizons | Not measured | 35.8% | Only Qwen2.5 72B Instruct measured |
| BALROG | Not measured | 16.2% | Only Qwen2.5 72B Instruct measured |
| TheAgentCompany | Not measured | 5.7% | Only Qwen2.5 72B Instruct measured |
| Vending-Bench 2 | $35 | Not measured | Only Grok 4.3 measured |
| LMArena Search | 1166 | Not measured | Only Grok 4.3 measured |
| Neither model has a result on the other 22 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index60
| Benchmark | Grok 4.3 | Qwen2.5 72B Instruct | Leader |
|---|---|---|---|
| GPQA Diamond | 91.4% | Not measured | Only Grok 4.3 measured |
| MMMU Pro (Vals AI) | 83.1% | Not measured | Only Grok 4.3 measured |
| GPQA Diamond (Artificial Analysis run) | 90.1% | 49.1% | Grok 4.3 |
| Humanity's Last Exam (Artificial Analysis run) | 37.2% | 3.6% | Grok 4.3 |
| GPQA Diamond (Epoch AI run) | 88.8% | 49.1% | Grok 4.3 |
| Chess puzzles | 18.0% | Not measured | Only Grok 4.3 measured |
| BIG-Bench Hard | Not measured | 79.8% | Only Qwen2.5 72B Instruct measured |
| LMArena Text: hard prompts | 1396 | 1271 | Grok 4.3 |
| LMArena Text: expert | 1384 | 1245 | Grok 4.3 |
| LMArena Text: hard prompts in English | 1402 | 1279 | Grok 4.3 |
| LMArena Vision: diagrams | 1241 | Not measured | Only Grok 4.3 measured |
| LMArena Vision: homework | 1241 | Not measured | Only Grok 4.3 measured |
| Neither model has a result on the other 10 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index30
| Benchmark | Grok 4.3 | Qwen2.5 72B Instruct | Leader |
|---|---|---|---|
| AA Math Index | Not measured | 14.0 | Only Qwen2.5 72B Instruct measured |
| AIME 2025 (Artificial Analysis run) | Not measured | 14.0% | Only Qwen2.5 72B Instruct measured |
| OTIS Mock AIME | 93.3% | 8.1% | Grok 4.3 |
| FrontierMath, tiers 1 to 3 | 42.8% | Not measured | Only Grok 4.3 measured |
| FrontierMath, tier 4 | 14.6% | Not measured | Only Grok 4.3 measured |
| ProofBench | 11.0% | Not measured | Only Grok 4.3 measured |
| MATH level 5 | Not measured | 63.2% | Only Qwen2.5 72B Instruct measured |
| LMArena Text: math | 1387 | 1283 | Grok 4.3 |
| LMArena Text: mathematical work | 1396 | 1290 | Grok 4.3 |
| Neither model has a result on the other 4 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index40
| Benchmark | Grok 4.3 | Qwen2.5 72B Instruct | Leader |
|---|---|---|---|
| MMLU Pro | 85.8% | Not measured | Only Grok 4.3 measured |
| LegalBench (Vals AI) | 84.5% | Not measured | Only Grok 4.3 measured |
| CaseLaw (Vals AI) | 79.3% | Not measured | Only Grok 4.3 measured |
| CorpFin (Vals AI) | 68.5% | Not measured | Only Grok 4.3 measured |
| TaxEval (Vals AI) | 70.8% | Not measured | Only Grok 4.3 measured |
| MortgageTax (Vals AI) | 48.3% | Not measured | Only Grok 4.3 measured |
| MedCode (Vals AI) | 38.1% | Not measured | Only Grok 4.3 measured |
| MMLU Pro (Artificial Analysis run) | Not measured | 72.0% | Only Qwen2.5 72B Instruct measured |
| SimpleQA Verified | 33.2% | Not measured | Only Grok 4.3 measured |
| MMLU | Not measured | 85.3% | Only Qwen2.5 72B Instruct measured |
| AI2 Reasoning Challenge | Not measured | 94.5% | Only Qwen2.5 72B Instruct measured |
| LMArena Text: science | 1399 | 1275 | Grok 4.3 |
| LMArena Text: medicine and healthcare | 1384 | 1258 | Grok 4.3 |
| LMArena Text: legal and government | 1386 | 1288 | Grok 4.3 |
| LMArena Text: business and finance | 1381 | 1253 | Grok 4.3 |
| Neither model has a result on the other 3 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
| Benchmark | Grok 4.3 | Qwen2.5 72B Instruct | Leader |
|---|---|---|---|
| Design Arena: websites | 1196 | Not measured | Only Grok 4.3 measured |
| Design Arena: UI components | 1192 | Not measured | Only Grok 4.3 measured |
| Design Arena: game dev | 1174 | Not measured | Only Grok 4.3 measured |
| Design Arena: data viz | 1179 | Not measured | Only Grok 4.3 measured |
| Design Arena: 3D | 1125 | Not measured | Only Grok 4.3 measured |
| Design Arena: all code categories | 1187 | Not measured | Only Grok 4.3 measured |
| Design Arena: SVG | 1081 | Not measured | Only Grok 4.3 measured |
| Design Arena: ASCII art | 1125 | Not measured | Only Grok 4.3 measured |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index190
| Benchmark | Grok 4.3 | Qwen2.5 72B Instruct | Leader |
|---|---|---|---|
| Public Benefits Bench (Vals AI) | 51.7% | Not measured | Only Grok 4.3 measured |
| SAGE (Vals AI) | 19.7% | Not measured | Only Grok 4.3 measured |
| MedScribe (Vals AI) | 74.4% | Not measured | Only Grok 4.3 measured |
| Vals Multimodal Index | 43.3% | Not measured | Only Grok 4.3 measured |
| IFBench | 83.3% | 36.9% | Grok 4.3 |
| Long context reasoning (AA-LCR) | 75.0% | Not measured | Only Grok 4.3 measured |
| DTBench | 90.7% | 62.9% | Grok 4.3 |
| LMCA | 38.3% | 13.4% | Grok 4.3 |
| HellaSwag | Not measured | 84.8% | Only Qwen2.5 72B Instruct measured |
| WinoGrande | Not measured | 82.3% | Only Qwen2.5 72B Instruct measured |
| PIQA | Not measured | 82.6% | Only Qwen2.5 72B Instruct measured |
| LMArena Text | 1397 | 1269 | Grok 4.3 |
| LMArena Text: instruction following | 1366 | 1254 | Grok 4.3 |
| LMArena Text: creative writing | 1380 | 1221 | Grok 4.3 |
| LMArena Vision | 1229 | Not measured | Only Grok 4.3 measured |
| LMArena Text: English | 1404 | 1283 | Grok 4.3 |
| LMArena Text: non-English | 1385 | 1252 | Grok 4.3 |
| LMArena Text: Chinese | 1422 | 1273 | Grok 4.3 |
| LMArena Text: German | 1395 | 1234 | Grok 4.3 |
| LMArena Text: Japanese | 1379 | 1180 | Grok 4.3 |
| LMArena Text: Korean | 1354 | 1188 | Grok 4.3 |
| LMArena Text: French | 1411 | 1280 | Grok 4.3 |
| LMArena Text: Spanish | 1399 | 1255 | Grok 4.3 |
| LMArena Text: Russian | 1399 | 1264 | Grok 4.3 |
| LMArena Text: Polish | 1398 | Not measured | Only Grok 4.3 measured |
| LMArena Text: multi-turn | 1406 | 1272 | Grok 4.3 |
| LMArena Text: longer queries | 1393 | 1281 | Grok 4.3 |
| LMArena Text: writing and language | 1385 | 1247 | Grok 4.3 |
| LMArena Text: entertainment and media | 1370 | 1211 | Grok 4.3 |
| LMArena Vision: English | 1235 | Not measured | Only Grok 4.3 measured |
| LMArena Vision: Chinese | 1257 | Not measured | Only Grok 4.3 measured |
| LMArena Vision: reading text | 1235 | Not measured | Only Grok 4.3 measured |
| LMArena Vision: humor | 1222 | Not measured | Only Grok 4.3 measured |
| LMArena Vision: recognizing things | 1217 | Not measured | Only Grok 4.3 measured |
| LMArena Vision: creative writing | 1237 | Not measured | Only Grok 4.3 measured |
| Neither model has a result on the other 11 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Grok 4.3 | Qwen2.5 72B Instruct | 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
Qwen2.5 72B Instruct costs less
Repository review
50K input + 3K output tokens
Qwen2.5 72B Instruct costs less
Agent loop
200K input + 10K output tokens
Qwen2.5 72B Instruct 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 | SpaceXAI Grok 4.3 | Qwen2.5 72B Instruct |
|---|---|---|
| Maker | xAI | Qwen |
| Weights | No public weights listed | Open weights |
| Released | April 30, 2026 | September 19, 2024 |
| Context window | 1M | 33K |
| Input $/Mtok | $1.25 | $0.36 |
| Output $/Mtok | $2.50 | $0.40 |
| Coding index | 37.4 | Not benchmarked |
| Independent measurements behind it | 3 | 0 |
| Cost per solved task | $0.31 | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Grok 4.3 better than Qwen2.5 72B Instruct for coding?
Only Grok 4.3 has a coding score, 37.4 on the Vibe Coding Index; Qwen2.5 72B Instruct has none, so the two cannot be ranked on coding. Pick Qwen2.5 72B Instruct for output tokens 6.3x cheaper ($0.40 vs $2.50 per million).
Which is cheaper, Grok 4.3 or Qwen2.5 72B Instruct?
Grok 4.3 costs $1.25 per million input tokens and $2.50 per million output tokens; Qwen2.5 72B Instruct costs $0.36 and $0.40. Cost per solved task is known for Grok 4.3 only.
Which has the bigger context window, Grok 4.3 or Qwen2.5 72B Instruct?
Grok 4.3, with 1M tokens against 33K.
How many benchmarks is the Grok 4.3 vs Qwen2.5 72B Instruct comparison based on?
38 benchmarks have a result for both models as of October 21, 2026. Grok 4.3's coding index rests on 3 independent benchmarks and Qwen2.5 72B Instruct's on 0. Where only one model has a result, the page says so instead of naming a winner.