Claude Fable 5.2 vs Gemma 4 31B
Claude Fable 5.2 leads Gemma 4 31B on coding, 87.8 to 21.9 on the Vibe Coding Index, and Gemma 4 31B costs 100.0x less per output token ($0.40 vs $40.00 per million). Updated November 14, 2026; 28 benchmarks have a result for both.
vs boards won of 28 shared
Claude Fable 5.2 or Gemma 4 31B: which to pick
Pick Claude Fable 5.2 for coding: it scores 87.8 on the Vibe Coding Index against 21.9 for Gemma 4 31B.
Pick Gemma 4 31B for output tokens 100.0x cheaper ($0.40 vs $40.00 per million).
Model AClaude Fable 5.2Anthropic87.8/1004 independent benchmarksNo cost per solved taskModel BGemma 4 31BGoogle21.9/1002 independent benchmarksNo cost per solved task
- Shared results
- 28
- Claude Fable 5.2 only
- 40
- Gemma 4 31B only
- 51
- Like-for-like categories
- 8 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
Claude Fable 5.2
Claude Fable 5.2 scores 87.8 against 21.9.
4 and 2 benchmarks behind the scores
Intelligence
AA Intelligence Index
Claude Fable 5.2
Claude Fable 5.2 leads, 64.6 against 4.4.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Gemma 4 31B has a AA Agentic Index result, 4.2.
Directional only
Web design
Design Arena: websites
No call
Only Claude Fable 5.2 has a Design Arena: websites result, 1351.
Directional only
General
Vals Index
No call
Only Claude Fable 5.2 has a Vals Index result, 71.3%.
Directional only
Long documents
Listed context window
Claude Fable 5.2
Claude Fable 5.2 takes 1M tokens against 262K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Gemma 4 31B
Claude Fable 5.2 $0.028, Gemma 4 31B $0.0003 per turn.
List prices
At a glance
- Coding index
- Claude Fable 5.2 87.8 from 4 benchmarks, Gemma 4 31B 21.9 from 2 benchmarks.
- Widest category
- Agentic work, 72.6 points apart on average over 3 shared benchmarks.
- Widest single gap
- LiveBench browser: Claude Fable 5.2 86.9% against Gemma 4 31B 0.0%, 86.9 points apart on a 0 to 100 scale.
- Output price
- Claude Fable 5.2 $40.00 against Gemma 4 31B $0.40 per million output tokens, 100.0x apart.
- Context window
- Claude Fable 5.2 takes 1M tokens, Gemma 4 31B 262K.
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 | Claude Fable 5.2 | Bars | Gemma 4 31B | Ties |
|---|---|---|---|---|
| Coding | 9 | 0 | ||
| Intelligence | 2 | 0 | ||
| Agentic work | 3 | 0 | ||
| Reasoning | 5 | 0 | ||
| Math | 1 | 0 | ||
| Knowledge | 2 | 0 | ||
| Web design | 1 | 0 | ||
| General | 5 | 0 | ||
| All shared boards | 28 | 0 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index90
| Benchmark | Claude Fable 5.2 | Gemma 4 31B | Leader |
|---|---|---|---|
| LiveBench coding | 79.2% | 20.3% | Claude Fable 5.2 |
| LMArena WebDev | 1838 | 1370 | Claude Fable 5.2 |
| AA Coding Index | Not measured | 43.4 | Only Gemma 4 31B measured |
| Terminal-Bench 4 (Vals AI) | 65.2% | Not measured | Only Claude Fable 5.2 measured |
| Vibe Code Bench (Vals AI) | 93.8% | Not measured | Only Claude Fable 5.2 measured |
| ProgramBench (Vals AI) | 19.0% | Not measured | Only Claude Fable 5.2 measured |
| SWE-bench Multimodal (Vals AI) | 62.3% | Not measured | Only Claude Fable 5.2 measured |
| SciCode | 67.7% | 45.5% | Claude Fable 5.2 |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 36.4% | Only Gemma 4 31B measured |
| Terminal-Bench 4.0 (Artificial Analysis run) | 61.6% | 0.0% | Claude Fable 5.2 |
| FrontierCode 2.0 | 49.1% | 0.0% | Claude Fable 5.2 |
| FrontierSWE | 69.9% | Not measured | Only Claude Fable 5.2 measured |
| WeirdML | Not measured | 52.3% | Only Gemma 4 31B measured |
| ALE-Bench | 2509.8 | 925.5 | Claude Fable 5.2 |
| LMArena Text: coding | Not measured | 1459 | Only Gemma 4 31B measured |
| LMArena Text: software and IT | Not measured | 1463 | Only Gemma 4 31B measured |
| LMArena WebDev: React | 1915 | 1362 | Claude Fable 5.2 |
| LMArena WebDev: HTML | 1738 | 1380 | Claude Fable 5.2 |
| SciCode | Not measured | 8.9% | Only Gemma 4 31B measured |
| GameBench 2 | 99.1 | 35.1 | Claude Fable 5.2 |
| CursorBench 4.0 | 60.0% | Not measured | Only Claude Fable 5.2 measured |
| Neither model has a result on the other 31 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Claude Fable 5.2 | Gemma 4 31B | Leader |
|---|---|---|---|
| AA Intelligence Index | 64.6 | 4.4 | Claude Fable 5.2 |
| Epoch Capabilities Index | 173.4 | 144.3 | Claude Fable 5.2 |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index30
| Benchmark | Claude Fable 5.2 | Gemma 4 31B | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 4.2 | Only Gemma 4 31B measured |
| Design Arena: agents building full-stack apps | 1402 | Not measured | Only Claude Fable 5.2 measured |
| Design Arena: agents building mobile apps | 1325 | Not measured | Only Claude Fable 5.2 measured |
| Design Arena: spreadsheets | 1434 | Not measured | Only Claude Fable 5.2 measured |
| LiveBench research | 94.5% | 38.1% | Claude Fable 5.2 |
| LiveBench browser | 86.9% | 0.0% | Claude Fable 5.2 |
| LiveBench long agentic | 82.9% | 8.5% | Claude Fable 5.2 |
| Tax Agent Bench (Vals AI) | 32.1% | Not measured | Only Claude Fable 5.2 measured |
| Legal Research Bench (Vals AI) | 46.1% | Not measured | Only Claude Fable 5.2 measured |
| Harvey's Legal Agent Benchmark (Vals AI) | 5.0% | Not measured | Only Claude Fable 5.2 measured |
| BioMysteryBench (Vals AI) | 83.7% | Not measured | Only Claude Fable 5.2 measured |
| Terminal-Bench-Science (Vals AI) | 55.7% | Not measured | Only Claude Fable 5.2 measured |
| Vals Science Index | 63.7% | Not measured | Only Claude Fable 5.2 measured |
| τ²-Bench Telecom | Not measured | 65.5% | Only Gemma 4 31B measured |
| APEX-Agents | 80.1% | Not measured | Only Claude Fable 5.2 measured |
| Remote Labor Index | 41.3% | Not measured | Only Claude Fable 5.2 measured |
| SpireClimb | 15.5% | Not measured | Only Claude Fable 5.2 measured |
| Neither model has a result on the other 27 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index50
| Benchmark | Claude Fable 5.2 | Gemma 4 31B | Leader |
|---|---|---|---|
| LiveBench reasoning | 88.8% | 68.3% | Claude Fable 5.2 |
| MysteryMechanism (Vals AI) | 51.8% | Not measured | Only Claude Fable 5.2 measured |
| GPQA Diamond (Artificial Analysis run) | Not measured | 85.7% | Only Gemma 4 31B measured |
| Humanity's Last Exam (Artificial Analysis run) | 74.3% | 23.6% | Claude Fable 5.2 |
| HLE-Diamond (Artificial Analysis run) | 80.2% | 7.2% | Claude Fable 5.2 |
| CritPt v2, hard subset (Artificial Analysis run) | 76.5% | 4.0% | Claude Fable 5.2 |
| GPQA Diamond (Epoch AI run) | 93.9% | 75.8% | Claude Fable 5.2 |
| ARC-AGI | 98.5% | Not measured | Only Claude Fable 5.2 measured |
| ARC-AGI-2 | 93.8% | Not measured | Only Claude Fable 5.2 measured |
| SimpleBench | 94.9% | Not measured | Only Claude Fable 5.2 measured |
| Chess puzzles | 79.0% | Not measured | Only Claude Fable 5.2 measured |
| Mystery game puzzles | 90.0% | Not measured | Only Claude Fable 5.2 measured |
| EBR-bench | 90.5% | Not measured | Only Claude Fable 5.2 measured |
| LMArena Text: hard prompts | Not measured | 1447 | Only Gemma 4 31B measured |
| LMArena Text: expert | Not measured | 1465 | Only Gemma 4 31B measured |
| LMArena Text: hard prompts in English | Not measured | 1456 | Only Gemma 4 31B measured |
| LMArena Vision: diagrams | Not measured | 1286 | Only Gemma 4 31B measured |
| LMArena Vision: homework | Not measured | 1310 | Only Gemma 4 31B measured |
| Kagi LLM Benchmark | Not measured | 63.5% | Only Gemma 4 31B measured |
| Humanity's Last Exam (Scale run) | 72.4% | Not measured | Only Claude Fable 5.2 measured |
| GoBench | 2566 | Not measured | Only Claude Fable 5.2 measured |
| Neither model has a result on the other 5 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Claude Fable 5.2 | Gemma 4 31B | Leader |
|---|---|---|---|
| OTIS Mock AIME | 99.7% | 73.3% | Claude Fable 5.2 |
| FrontierMath, tiers 1 to 3 | 95.1% | Not measured | Only Claude Fable 5.2 measured |
| FrontierMath, tier 4 | 100.0% | Not measured | Only Claude Fable 5.2 measured |
| ProofBench | 100.0% | Not measured | Only Claude Fable 5.2 measured |
| LMArena Text: math | Not measured | 1462 | Only Gemma 4 31B measured |
| LMArena Text: mathematical work | Not measured | 1464 | Only Gemma 4 31B measured |
| MathArena | 82.8% | Not measured | Only Claude Fable 5.2 measured |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Claude Fable 5.2 | Gemma 4 31B | Leader |
|---|---|---|---|
| CaseLaw (Vals AI) | Not measured | 52.6% | Only Gemma 4 31B measured |
| MortgageTax (Vals AI) | Not measured | 61.4% | Only Gemma 4 31B measured |
| AA-Omniscience | 66.8 | -48.5 | Claude Fable 5.2 |
| SimpleQA Verified | 94.0% | 10.4% | Claude Fable 5.2 |
| LMArena Text: science | Not measured | 1456 | Only Gemma 4 31B measured |
| LMArena Text: medicine and healthcare | Not measured | 1450 | Only Gemma 4 31B measured |
| LMArena Text: legal and government | Not measured | 1438 | Only Gemma 4 31B measured |
| LMArena Text: business and finance | Not measured | 1431 | Only Gemma 4 31B measured |
| FACTS Benchmark Suite | Not measured | 50.1% | Only Gemma 4 31B measured |
| PRBench Finance | 47.1% | Not measured | Only Claude Fable 5.2 measured |
| PRBench Legal | 49.0% | Not measured | Only Claude Fable 5.2 measured |
| Neither model has a result on the other 10 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Claude Fable 5.2 | Gemma 4 31B | Leader |
|---|---|---|---|
| Design Arena: websites | 1351 | Not measured | Only Claude Fable 5.2 measured |
| Design Arena: UI components | 1371 | Not measured | Only Claude Fable 5.2 measured |
| Design Arena: game dev | 1396 | Not measured | Only Claude Fable 5.2 measured |
| Design Arena: 3D | 1517 | Not measured | Only Claude Fable 5.2 measured |
| Design Arena: all code categories | 1373 | Not measured | Only Claude Fable 5.2 measured |
| LiveBench design | 85.1% | 56.5% | Claude Fable 5.2 |
| Neither model has a result on the other 3 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index50
| Benchmark | Claude Fable 5.2 | Gemma 4 31B | Leader |
|---|---|---|---|
| Vals Index | 71.3% | Not measured | Only Claude Fable 5.2 measured |
| LiveBench data analysis | 81.5% | 53.3% | Claude Fable 5.2 |
| SAGE (Vals AI) | Not measured | 82.8% | Only Gemma 4 31B measured |
| IFBench | Not measured | 75.6% | Only Gemma 4 31B measured |
| Long context reasoning (AA-LCR) | 83.3% | 69.7% | Claude Fable 5.2 |
| GDPval-AA | 1831 | 627 | Claude Fable 5.2 |
| DTBench | 99.2% | 82.7% | Claude Fable 5.2 |
| LMCA | 70.6% | 39.3% | Claude Fable 5.2 |
| Surface Evolver Bench | Not measured | 30.6% | Only Gemma 4 31B measured |
| Furniture assembly | 90.0% | Not measured | Only Claude Fable 5.2 measured |
| LMArena Text | Not measured | 1443 | Only Gemma 4 31B measured |
| LMArena Text: instruction following | Not measured | 1433 | Only Gemma 4 31B measured |
| LMArena Text: creative writing | Not measured | 1415 | Only Gemma 4 31B measured |
| LMArena Vision | Not measured | 1279 | Only Gemma 4 31B measured |
| LMArena Text: English | Not measured | 1451 | Only Gemma 4 31B measured |
| LMArena Text: non-English | Not measured | 1430 | Only Gemma 4 31B measured |
| LMArena Text: Chinese | Not measured | 1476 | Only Gemma 4 31B measured |
| LMArena Text: French | Not measured | 1434 | Only Gemma 4 31B measured |
| LMArena Text: Spanish | Not measured | 1444 | Only Gemma 4 31B measured |
| LMArena Text: Russian | Not measured | 1460 | Only Gemma 4 31B measured |
| LMArena Text: Polish | Not measured | 1431 | Only Gemma 4 31B measured |
| LMArena Text: multi-turn | Not measured | 1452 | Only Gemma 4 31B measured |
| LMArena Text: longer queries | Not measured | 1445 | Only Gemma 4 31B measured |
| LMArena Text: writing and language | Not measured | 1431 | Only Gemma 4 31B measured |
| LMArena Text: entertainment and media | Not measured | 1405 | Only Gemma 4 31B measured |
| LMArena Vision: English | Not measured | 1272 | Only Gemma 4 31B measured |
| LMArena Vision: Chinese | Not measured | 1327 | Only Gemma 4 31B measured |
| LMArena Vision: reading text | Not measured | 1285 | Only Gemma 4 31B measured |
| LMArena Vision: humor | Not measured | 1280 | Only Gemma 4 31B measured |
| LMArena Vision: captioning | Not measured | 1209 | Only Gemma 4 31B measured |
| LMArena Vision: recognizing things | Not measured | 1227 | Only Gemma 4 31B measured |
| LMArena Vision: creative writing | Not measured | 1286 | Only Gemma 4 31B measured |
| LMArena Document | Not measured | 1434 | Only Gemma 4 31B measured |
| EQ-Bench 4 | Not measured | 1110 | Only Gemma 4 31B measured |
| Neither model has a result on the other 15 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Claude Fable 5.2 | Gemma 4 31B | 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
Claude Fable 5.2$0.028Fits in one request
Gemma 4 31B$0.0003Fits in one request
Gemma 4 31B costs less
Repository review
50K input + 3K output tokens
Claude Fable 5.2$0.520Fits in one request
Gemma 4 31B$0.0082Fits in one request
Gemma 4 31B costs less
Agent loop
200K input + 10K output tokens
Claude Fable 5.2$2.00Fits in one request
Gemma 4 31B$0.032Fits in one request
Gemma 4 31B 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 | Anthropic Claude Fable 5.2 | Google Gemma 4 31B |
|---|---|---|
| Maker | Anthropic | |
| Weights | No public weights listed | Open weights |
| Released | November 9, 2026 | April 2, 2026 |
| Context window | 1M | 262K |
| Input $/Mtok | $8.00 | $0.14 |
| Output $/Mtok | $40.00 | $0.40 |
| Coding index | 87.8 | 21.9 |
| Independent measurements behind it | 4 | 2 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Claude Fable 5.2 better than Gemma 4 31B for coding?
Pick Claude Fable 5.2 for coding: it scores 87.8 on the Vibe Coding Index against 21.9 for Gemma 4 31B. Pick Gemma 4 31B for output tokens 100.0x cheaper ($0.40 vs $40.00 per million).
Which is cheaper, Claude Fable 5.2 or Gemma 4 31B?
Claude Fable 5.2 costs $8.00 per million input tokens and $40.00 per million output tokens; Gemma 4 31B costs $0.14 and $0.40. Cost per solved task is known for neither of them.
Which has the bigger context window, Claude Fable 5.2 or Gemma 4 31B?
Claude Fable 5.2, with 1M tokens against 262K.
How many benchmarks is the Claude Fable 5.2 vs Gemma 4 31B comparison based on?
28 benchmarks have a result for both models as of November 14, 2026. Claude Fable 5.2's coding index rests on 4 independent benchmarks and Gemma 4 31B's on 2. Where only one model has a result, the page says so instead of naming a winner.