Claude Opus 4.7 vs Inkling Small
Claude Opus 4.7 leads Inkling Small on coding, 67.5 to 27.3 on the Vibe Coding Index, and Inkling Small costs 20.8x less per output token ($1.20 vs $25.00 per million). Updated October 22, 2026; 76 benchmarks have a result for both.
Claude Opus 4.7 or Inkling Small: which to pick
Pick Claude Opus 4.7 for coding: it scores 67.5 on the Vibe Coding Index against 27.3 for Inkling Small.
Pick Inkling Small for output tokens 20.8x cheaper ($1.20 vs $25.00 per million) and a lower cost per solved task ($0.16 against $2.95).
- Shared results
- 76
- Claude Opus 4.7 only
- 27
- Inkling Small only
- 8
- 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
Claude Opus 4.7
Claude Opus 4.7 scores 67.5 against 27.3.
4 benchmarks behind each score
Intelligence
AA Intelligence Index
Claude Opus 4.7
Claude Opus 4.7 leads, 40.7 against 25.7.
Same board, both measured
Agentic work
AA Agentic Index
Claude Opus 4.7
Claude Opus 4.7 leads, 38.6 against 23.5.
Same board, both measured
Reasoning
GPQA Diamond
Claude Opus 4.7
Claude Opus 4.7 leads, 90.2% against 83.6%.
Same board, both measured
Math
AIME
No call
Only Claude Opus 4.7 has a AIME result, 96.3%.
Directional only
Knowledge
MMLU Pro
Claude Opus 4.7
Claude Opus 4.7 leads, 89.9% against 85.6%.
Same board, both measured
General
Vals Index
No call
Only Inkling Small has a Vals Index result, 27.6%.
Directional only
Long documents
Listed context window
Claude Opus 4.7
Claude Opus 4.7 takes 1M tokens against 524K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Inkling Small
Claude Opus 4.7 $0.018, Inkling Small $0.0010 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
Inkling Small
Claude Opus 4.7 $2.95, Inkling Small $0.16.
Published attempt cost and pass rate
At a glance
- Coding index
- Claude Opus 4.7 67.5 from 4 benchmarks, Inkling Small 27.3 from 4 benchmarks.
- Widest category
- Math, 18.7 points apart on average over 6 shared benchmarks.
- Widest single gap
- Vibe Code Bench (Vals AI): Claude Opus 4.7 71.0% against Inkling Small 19.1%, 52.0 points apart on a 0 to 100 scale.
- Output price
- Claude Opus 4.7 $25.00 against Inkling Small $1.20 per million output tokens, 20.8x apart.
- Context window
- Claude Opus 4.7 takes 1M tokens, Inkling Small 524K.
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 Opus 4.7 | Bars | Inkling Small | Ties |
|---|---|---|---|---|
| Coding | 10 | 3 | ||
| Intelligence | 2 | 0 | ||
| Agentic work | 5 | 0 | ||
| Reasoning | 12 | 0 | ||
| Math | 6 | 0 | ||
| Knowledge | 9 | 2 | ||
| General | 26 | 1 | ||
| All shared boards | 70 | 6 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index103
| Benchmark | Claude Opus 4.7 | Inkling Small | Leader |
|---|---|---|---|
| LMArena WebDev | 1558 | 1409 | Claude Opus 4.7 |
| AA Coding Index | 73.6 | 52.9 | Claude Opus 4.7 |
| LiveCodeBench (Vals AI) | 85.1% | 85.9% | Inkling Small |
| SWE-bench Verified (Vals AI) | 82.0% | 82.2% | Inkling Small |
| Terminal-Bench 4 (Vals AI) | Not measured | 0.0% | Only Inkling Small measured |
| Vibe Code Bench (Vals AI) | 71.0% | 19.1% | Claude Opus 4.7 |
| Code migration (Vals AI) | 43.9% | 13.7% | Claude Opus 4.7 |
| ProgramBench (Vals AI) | 0.0% | 0.5% | Inkling Small |
| IOI (Vals AI) | Not measured | 9.3% | Only Inkling Small measured |
| Terminal-Bench 2.1 (Vals AI) | 68.5% | 55.1% | Claude Opus 4.7 |
| SkillsBench (Vals AI) | Not measured | 33.6% | Only Inkling Small measured |
| CyberBench (Vals AI) | Not measured | 64.1% | Only Inkling Small measured |
| SciCode | Not measured | 49.7% | Only Inkling Small measured |
| Terminal-Bench Hard (Artificial Analysis run) | 54.5% | Not measured | Only Claude Opus 4.7 measured |
| Terminal-Bench | 80.2% | Not measured | Only Claude Opus 4.7 measured |
| FrontierCode | 38.5% | Not measured | Only Claude Opus 4.7 measured |
| GSO-Bench | 44.1% | Not measured | Only Claude Opus 4.7 measured |
| WeirdML | 76.4% | Not measured | Only Claude Opus 4.7 measured |
| MirrorCode | 31.1% | Not measured | Only Claude Opus 4.7 measured |
| LMArena Text: coding | 1519 | 1450 | Claude Opus 4.7 |
| LMArena Text: software and IT | 1516 | 1447 | Claude Opus 4.7 |
| LMArena WebDev: React | 1560 | 1399 | Claude Opus 4.7 |
| LMArena WebDev: HTML | 1554 | 1398 | Claude Opus 4.7 |
| LMArena WebDev: from an image | 1577 | 1410 | Claude Opus 4.7 |
| Neither model has a result on the other 15 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Claude Opus 4.7 | Inkling Small | Leader |
|---|---|---|---|
| AA Intelligence Index | 40.7 | 25.7 | Claude Opus 4.7 |
| Epoch Capabilities Index | 156.8 | 150.3 | Claude Opus 4.7 |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index50
| Benchmark | Claude Opus 4.7 | Inkling Small | Leader |
|---|---|---|---|
| AA Agentic Index | 38.6 | 23.5 | Claude Opus 4.7 |
| Finance Agent (Vals AI) | 51.5% | 41.3% | Claude Opus 4.7 |
| Tax Agent Bench (Vals AI) | 23.7% | 15.4% | Claude Opus 4.7 |
| Legal Research Bench (Vals AI) | 38.5% | 25.5% | Claude Opus 4.7 |
| Harvey's Legal Agent Benchmark (Vals AI) | Not measured | 2.9% | Only Inkling Small measured |
| Excel Modeling Benchmark (Vals AI) | 63.7% | 31.8% | Claude Opus 4.7 |
| τ²-Bench Telecom | 88.6% | Not measured | Only Claude Opus 4.7 measured |
| APEX-Agents | 49.2% | Not measured | Only Claude Opus 4.7 measured |
| OSWorld 2.0 | 18.2% | Not measured | Only Claude Opus 4.7 measured |
| PostTrainBench | 28.6% | Not measured | Only Claude Opus 4.7 measured |
| Vending-Bench 2 | $10,937 | Not measured | Only Claude Opus 4.7 measured |
| LMArena Search | 1230 | Not measured | Only Claude Opus 4.7 measured |
| Neither model has a result on the other 29 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index120
| Benchmark | Claude Opus 4.7 | Inkling Small | Leader |
|---|---|---|---|
| GPQA Diamond | 90.2% | 83.6% | Claude Opus 4.7 |
| MMMU Pro (Vals AI) | 85.5% | Not measured | Only Claude Opus 4.7 measured |
| GPQA Diamond (Artificial Analysis run) | 91.4% | 89.5% | Claude Opus 4.7 |
| Humanity's Last Exam (Artificial Analysis run) | 42.3% | 33.3% | Claude Opus 4.7 |
| GPQA Diamond (Epoch AI run) | 90.2% | 88.5% | Claude Opus 4.7 |
| ARC-AGI | 93.5% | 84.0% | Claude Opus 4.7 |
| ARC-AGI-2 | 75.8% | 40.1% | Claude Opus 4.7 |
| SimpleBench | 61.7% | Not measured | Only Claude Opus 4.7 measured |
| Chess puzzles | 41.0% | Not measured | Only Claude Opus 4.7 measured |
| Mystery game puzzles | 28.0% | 6.0% | Claude Opus 4.7 |
| EBR-bench | 19.0% | Not measured | Only Claude Opus 4.7 measured |
| LMArena Text: hard prompts | 1506 | 1423 | Claude Opus 4.7 |
| LMArena Text: expert | 1521 | 1441 | Claude Opus 4.7 |
| LMArena Text: hard prompts in English | 1508 | 1429 | Claude Opus 4.7 |
| LMArena Vision: diagrams | 1336 | 1251 | Claude Opus 4.7 |
| LMArena Vision: homework | 1337 | 1257 | Claude Opus 4.7 |
| Kagi LLM Benchmark | 80.7% | Not measured | Only Claude Opus 4.7 measured |
| Neither model has a result on the other 5 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index60
| Benchmark | Claude Opus 4.7 | Inkling Small | Leader |
|---|---|---|---|
| AIME | 96.3% | Not measured | Only Claude Opus 4.7 measured |
| OTIS Mock AIME | 97.8% | 90.0% | Claude Opus 4.7 |
| FrontierMath, tiers 1 to 3 | 70.2% | 46.3% | Claude Opus 4.7 |
| FrontierMath, tier 4 | 31.7% | 17.1% | Claude Opus 4.7 |
| ProofBench | 54.0% | 6.0% | Claude Opus 4.7 |
| LMArena Text: math | 1498 | 1459 | Claude Opus 4.7 |
| LMArena Text: mathematical work | 1509 | 1441 | Claude Opus 4.7 |
| Neither model has a result on the other 6 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index92
| Benchmark | Claude Opus 4.7 | Inkling Small | Leader |
|---|---|---|---|
| MMLU Pro | 89.9% | 85.6% | Claude Opus 4.7 |
| LegalBench (Vals AI) | 85.3% | 83.0% | Claude Opus 4.7 |
| CaseLaw (Vals AI) | 68.4% | Not measured | Only Claude Opus 4.7 measured |
| CorpFin (Vals AI) | 66.1% | 69.6% | Inkling Small |
| TaxEval (Vals AI) | 75.3% | 75.5% | Inkling Small |
| MortgageTax (Vals AI) | 70.3% | 62.3% | Claude Opus 4.7 |
| MedCode (Vals AI) | 54.9% | 37.9% | Claude Opus 4.7 |
| SimpleQA Verified | 51.7% | 19.1% | Claude Opus 4.7 |
| LMArena Text: science | 1508 | 1428 | Claude Opus 4.7 |
| LMArena Text: medicine and healthcare | 1495 | 1413 | Claude Opus 4.7 |
| LMArena Text: legal and government | 1504 | 1415 | Claude Opus 4.7 |
| LMArena Text: business and finance | 1490 | 1413 | Claude Opus 4.7 |
| FACTS Benchmark Suite | 45.5% | Not measured | Only Claude Opus 4.7 measured |
| Neither model has a result on the other 5 boards here | |||
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Claude Opus 4.7 | Inkling Small | Leader |
|---|---|---|---|
| Neither model has a result on 9 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index261
| Benchmark | Claude Opus 4.7 | Inkling Small | Leader |
|---|---|---|---|
| Vals Index | Not measured | 27.6% | Only Inkling Small measured |
| Public Benefits Bench (Vals AI) | Not measured | 59.4% | Only Inkling Small measured |
| SAGE (Vals AI) | 56.1% | 32.7% | Claude Opus 4.7 |
| MedScribe (Vals AI) | 83.0% | 84.1% | Inkling Small |
| Vals Multimodal Index | 67.4% | 50.0% | Claude Opus 4.7 |
| IFBench | 58.6% | Not measured | Only Claude Opus 4.7 measured |
| Long context reasoning (AA-LCR) | 78.7% | 75.7% | Claude Opus 4.7 |
| DTBench | 94.7% | Not measured | Only Claude Opus 4.7 measured |
| LMCA | 52.2% | Not measured | Only Claude Opus 4.7 measured |
| Furniture assembly | 33.3% | Not measured | Only Claude Opus 4.7 measured |
| ExploitBench | 26.5% | Not measured | Only Claude Opus 4.7 measured |
| LMArena Text | 1490 | 1413 | Claude Opus 4.7 |
| LMArena Text: instruction following | 1498 | 1399 | Claude Opus 4.7 |
| LMArena Text: creative writing | 1486 | 1334 | Claude Opus 4.7 |
| LMArena Vision | 1317 | 1237 | Claude Opus 4.7 |
| LMArena Text: English | 1496 | 1421 | Claude Opus 4.7 |
| LMArena Text: non-English | 1480 | 1401 | Claude Opus 4.7 |
| LMArena Text: Chinese | 1531 | 1464 | Claude Opus 4.7 |
| LMArena Text: German | 1495 | 1408 | Claude Opus 4.7 |
| LMArena Text: Japanese | 1474 | 1401 | Claude Opus 4.7 |
| LMArena Text: Korean | 1464 | 1367 | Claude Opus 4.7 |
| LMArena Text: French | 1503 | 1435 | Claude Opus 4.7 |
| LMArena Text: Spanish | 1494 | 1422 | Claude Opus 4.7 |
| LMArena Text: Russian | 1494 | 1394 | Claude Opus 4.7 |
| LMArena Text: Polish | 1489 | 1428 | Claude Opus 4.7 |
| LMArena Text: multi-turn | 1505 | 1416 | Claude Opus 4.7 |
| LMArena Text: longer queries | 1505 | 1402 | Claude Opus 4.7 |
| LMArena Text: writing and language | 1492 | 1357 | Claude Opus 4.7 |
| LMArena Text: entertainment and media | 1474 | 1352 | Claude Opus 4.7 |
| LMArena Vision: English | 1322 | 1243 | Claude Opus 4.7 |
| LMArena Vision: Chinese | 1371 | 1292 | Claude Opus 4.7 |
| LMArena Vision: reading text | 1327 | 1243 | Claude Opus 4.7 |
| LMArena Vision: humor | 1309 | 1207 | Claude Opus 4.7 |
| LMArena Vision: recognizing things | 1278 | Not measured | Only Claude Opus 4.7 measured |
| LMArena Vision: creative writing | 1326 | 1205 | Claude Opus 4.7 |
| LMArena Document | 1495 | Not measured | Only Claude Opus 4.7 measured |
| Neither model has a result on the other 10 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Claude Opus 4.7 | Inkling Small | 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
Inkling Small costs less
Repository review
50K input + 3K output tokens
Inkling Small costs less
Agent loop
200K input + 10K output tokens
Inkling Small 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 Opus 4.7 | Thinking Machines Inkling Small |
|---|---|---|
| Maker | Anthropic | Thinking Machines |
| Weights | No public weights listed | Open weights |
| Released | April 16, 2026 | July 30, 2026 |
| Context window | 1M | 524K |
| Input $/Mtok | $5.00 | $0.45 |
| Output $/Mtok | $25.00 | $1.20 |
| Coding index | 67.5 | 27.3 |
| Independent measurements behind it | 4 | 4 |
| Cost per solved task | $2.95 | $0.16 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Claude Opus 4.7 better than Inkling Small for coding?
Pick Claude Opus 4.7 for coding: it scores 67.5 on the Vibe Coding Index against 27.3 for Inkling Small. Pick Inkling Small for output tokens 20.8x cheaper ($1.20 vs $25.00 per million) and a lower cost per solved task ($0.16 against $2.95).
Which is cheaper, Claude Opus 4.7 or Inkling Small?
Claude Opus 4.7 costs $5.00 per million input tokens and $25.00 per million output tokens; Inkling Small costs $0.45 and $1.20. Per solved SWE-bench Verified task, Claude Opus 4.7 costs $2.95 and Inkling Small $0.16.
Which has the bigger context window, Claude Opus 4.7 or Inkling Small?
Claude Opus 4.7, with 1M tokens against 524K.
How many benchmarks is the Claude Opus 4.7 vs Inkling Small comparison based on?
76 benchmarks have a result for both models as of October 22, 2026. Claude Opus 4.7's coding index rests on 4 independent benchmarks and Inkling Small's on 4. Where only one model has a result, the page says so instead of naming a winner.