Granite 4.2 8B vs Qwen3 32B
Granite 4.2 8B leads Qwen3 32B on the Artificial Analysis Intelligence Index, 11.1 to 7.3, and Granite 4.2 8B costs 2.0x less per output token ($0.20 vs $0.40 per million). Updated October 22, 2026; 25 benchmarks have a result for both.
Granite 4.2 8B or Qwen3 32B: which to pick
Neither model has a coding score yet, so the two cannot be ranked on coding.
Pick Granite 4.2 8B for output tokens 2.0x cheaper ($0.20 vs $0.40 per million) and work measured by SciCode or GPQA Diamond (Artificial Analysis run), where it leads.
- Shared results
- 25
- Granite 4.2 8B only
- 2
- Qwen3 32B only
- 20
- Like-for-like categories
- 5 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
Neither has a coding score yet.
One side not measured
Intelligence
AA Intelligence Index
Granite 4.2 8B
Granite 4.2 8B leads, 11.1 against 7.3.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Granite 4.2 8B has a AA Agentic Index result, 1.3.
Directional only
Long documents
Listed context window
No call
Both take 131K tokens.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Granite 4.2 8B
Granite 4.2 8B $0.0002, Qwen3 32B $0.0003 per turn.
List prices
At a glance
- Coding index
- Granite 4.2 8B not benchmarked, Qwen3 32B not benchmarked.
- Widest category
- General, 7.0 points apart on average over 12 shared benchmarks.
- Widest single gap
- Long context reasoning (AA-LCR): Granite 4.2 8B 45.0% against Qwen3 32B 0.0%, 45.0 points apart on a 0 to 100 scale.
- Output price
- Granite 4.2 8B $0.20 against Qwen3 32B $0.40 per million output tokens, 2.0x apart.
- Context window
- Both take 131K tokens.
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 | Granite 4.2 8B | Bars | Qwen3 32B | Ties |
|---|---|---|---|---|
| Coding | 2 | 1 | ||
| Intelligence | 1 | 0 | ||
| Reasoning | 2 | 3 | ||
| Knowledge | 2 | 2 | ||
| General | 2 | 10 | ||
| All shared boards | 9 | 16 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index21
| Benchmark | Granite 4.2 8B | Qwen3 32B | Leader |
|---|---|---|---|
| Aider polyglot | Not measured | 40.0% | Only Qwen3 32B measured |
| AA Coding Index | 22.4 | Not measured | Only Granite 4.2 8B measured |
| LiveCodeBench (Artificial Analysis run) | Not measured | 54.6% | Only Qwen3 32B measured |
| SciCode | 31.5% | 36.0% | Qwen3 32B |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 3.0% | Only Qwen3 32B measured |
| LMArena Text: coding | 1378 | 1357 | Granite 4.2 8B |
| LMArena Text: software and IT | 1367 | 1366 | Granite 4.2 8B |
| Neither model has a result on the other 36 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Granite 4.2 8B | Qwen3 32B | Leader |
|---|---|---|---|
| AA Intelligence Index | 11.1 | 7.3 | Granite 4.2 8B |
| Epoch Capabilities Index | Not measured | 138.4 | Only Qwen3 32B measured |
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index4 boards, none shared
| Benchmark | Granite 4.2 8B | Qwen3 32B | Leader |
|---|---|---|---|
| AA Agentic Index | 1.3 | Not measured | Only Granite 4.2 8B measured |
| τ²-Bench Telecom | Not measured | 29.8% | Only Qwen3 32B measured |
| Berkeley Function Calling Leaderboard (BFCL) | Not measured | 48.7% | Only Qwen3 32B measured |
| BrowseComp-Plus | Not measured | 10.7% | Only Qwen3 32B measured |
| Neither model has a result on the other 38 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index23
| Benchmark | Granite 4.2 8B | Qwen3 32B | Leader |
|---|---|---|---|
| GPQA Diamond (Artificial Analysis run) | 63.1% | 66.8% | Qwen3 32B |
| Humanity's Last Exam (Artificial Analysis run) | 9.7% | 7.4% | Granite 4.2 8B |
| GPQA Diamond (Epoch AI run) | Not measured | 65.7% | Only Qwen3 32B measured |
| LMArena Text: hard prompts | 1327 | 1334 | Qwen3 32B |
| LMArena Text: expert | 1386 | 1362 | Granite 4.2 8B |
| LMArena Text: hard prompts in English | 1345 | 1350 | Qwen3 32B |
| Kagi LLM Benchmark | Not measured | 48.7% | Only Qwen3 32B measured |
| Neither model has a result on the other 16 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index5 boards, none shared
| Benchmark | Granite 4.2 8B | Qwen3 32B | Leader |
|---|---|---|---|
| AA Math Index | Not measured | 73.0 | Only Qwen3 32B measured |
| AIME 2025 (Artificial Analysis run) | Not measured | 73.0% | Only Qwen3 32B measured |
| OTIS Mock AIME | Not measured | 66.9% | Only Qwen3 32B measured |
| LMArena Text: math | Not measured | 1399 | Only Qwen3 32B measured |
| LMArena Text: mathematical work | Not measured | 1420 | Only Qwen3 32B measured |
| Neither model has a result on the other 9 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index22
| Benchmark | Granite 4.2 8B | Qwen3 32B | Leader |
|---|---|---|---|
| MMLU Pro (Artificial Analysis run) | Not measured | 79.8% | Only Qwen3 32B measured |
| LMArena Text: science | 1336 | 1372 | Qwen3 32B |
| LMArena Text: medicine and healthcare | 1336 | 1368 | Qwen3 32B |
| LMArena Text: legal and government | 1327 | 1324 | Granite 4.2 8B |
| LMArena Text: business and finance | 1322 | 1318 | Granite 4.2 8B |
| Neither model has a result on the other 15 boards here | |||
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Granite 4.2 8B | Qwen3 32B | Leader |
|---|---|---|---|
| Neither model has a result on 9 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index210
| Benchmark | Granite 4.2 8B | Qwen3 32B | Leader |
|---|---|---|---|
| IFBench | Not measured | 36.3% | Only Qwen3 32B measured |
| Long context reasoning (AA-LCR) | 45.0% | 0.0% | Granite 4.2 8B |
| DTBench | Not measured | 67.5% | Only Qwen3 32B measured |
| LMCA | Not measured | 17.3% | Only Qwen3 32B measured |
| Fiction.LiveBench | Not measured | 74.2% | Only Qwen3 32B measured |
| LMArena Text | 1319 | 1340 | Qwen3 32B |
| LMArena Text: instruction following | 1300 | 1305 | Qwen3 32B |
| LMArena Text: creative writing | 1236 | 1297 | Qwen3 32B |
| LMArena Text: English | 1337 | 1359 | Qwen3 32B |
| LMArena Text: non-English | 1301 | 1317 | Qwen3 32B |
| LMArena Text: Chinese | 1367 | 1357 | Granite 4.2 8B |
| LMArena Text: German | Not measured | 1341 | Only Qwen3 32B measured |
| LMArena Text: Russian | 1282 | 1311 | Qwen3 32B |
| LMArena Text: multi-turn | 1301 | 1331 | Qwen3 32B |
| LMArena Text: longer queries | 1325 | 1327 | Qwen3 32B |
| LMArena Text: writing and language | 1269 | 1298 | Qwen3 32B |
| LMArena Text: entertainment and media | 1271 | 1279 | Qwen3 32B |
| Neither model has a result on the other 31 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Granite 4.2 8B | Qwen3 32B | 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
Granite 4.2 8B costs less
Repository review
50K input + 3K output tokens
Granite 4.2 8B costs less
Agent loop
200K input + 10K output tokens
Granite 4.2 8B 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 | IBM Granite 4.2 8B | Qwen3 32B |
|---|---|---|
| Maker | IBM Granite | Qwen |
| Weights | Open weights | Open weights |
| Released | August 31, 2026 | April 28, 2025 |
| Context window | 131K | 131K |
| Input $/Mtok | $0.08 | $0.14 |
| Output $/Mtok | $0.20 | $0.40 |
| 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 Granite 4.2 8B better than Qwen3 32B for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Granite 4.2 8B for output tokens 2.0x cheaper ($0.20 vs $0.40 per million) and work measured by SciCode or GPQA Diamond (Artificial Analysis run), where it leads.
Which is cheaper, Granite 4.2 8B or Qwen3 32B?
Granite 4.2 8B costs $0.08 per million input tokens and $0.20 per million output tokens; Qwen3 32B costs $0.14 and $0.40. Cost per solved task is known for neither of them.
Which has the bigger context window, Granite 4.2 8B or Qwen3 32B?
Neither: both take 131K tokens.
How many benchmarks is the Granite 4.2 8B vs Qwen3 32B comparison based on?
25 benchmarks have a result for both models as of October 22, 2026. Granite 4.2 8B's coding index rests on 0 independent benchmarks and Qwen3 32B's on 0. Where only one model has a result, the page says so instead of naming a winner.