Claude Sonnet 5.5 vs GPT-5.4 Mini
Claude Sonnet 5.5 leads GPT-5.4 Mini on coding, 87.3 to 35.8 on the Vibe Coding Index, and GPT-5.4 Mini costs 2.2x less per output token ($4.50 vs $10.00 per million). Updated October 22, 2026; 60 benchmarks have a result for both.
Claude Sonnet 5.5 or GPT-5.4 Mini: which to pick
Pick Claude Sonnet 5.5 for coding: it scores 87.3 on the Vibe Coding Index against 35.8 for GPT-5.4 Mini.
Pick GPT-5.4 Mini for output tokens 2.2x cheaper ($4.50 vs $10.00 per million).
- Shared results
- 60
- Claude Sonnet 5.5 only
- 36
- GPT-5.4 Mini only
- 29
- 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 Sonnet 5.5
Claude Sonnet 5.5 scores 87.3 against 35.8.
5 benchmarks behind each score
Intelligence
AA Intelligence Index
Claude Sonnet 5.5
Claude Sonnet 5.5 leads, 56.0 against 24.1.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only GPT-5.4 Mini has a AA Agentic Index result, 17.9.
Directional only
Reasoning
GPQA Diamond
No call
Only GPT-5.4 Mini has a GPQA Diamond result, 83.1%.
Directional only
Math
AIME
No call
Only GPT-5.4 Mini has a AIME result, 95.6%.
Directional only
Knowledge
MMLU Pro
No call
Only GPT-5.4 Mini has a MMLU Pro result, 84.5%.
Directional only
Web design
Design Arena: websites
No call
Only Claude Sonnet 5.5 has a Design Arena: websites result, 1326.
Directional only
General
Vals Index
No call
Only Claude Sonnet 5.5 has a Vals Index result, 70.8%.
Directional only
Long documents
Listed context window
Claude Sonnet 5.5
Claude Sonnet 5.5 takes 1M tokens against 400K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
GPT-5.4 Mini
Claude Sonnet 5.5 $0.0070, GPT-5.4 Mini $0.0030 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only GPT-5.4 Mini has the number, $0.70.
Directional only
At a glance
- Coding index
- Claude Sonnet 5.5 87.3 from 5 benchmarks, GPT-5.4 Mini 35.8 from 5 benchmarks.
- Widest category
- Math, 37.5 points apart on average over 6 shared benchmarks.
- Widest single gap
- Terminal-Bench 4 (Vals AI): Claude Sonnet 5.5 64.1% against GPT-5.4 Mini 2.5%, 82.2 points apart on a 0 to 100 scale.
- Output price
- Claude Sonnet 5.5 $10.00 against GPT-5.4 Mini $4.50 per million output tokens, 2.2x apart.
- Context window
- Claude Sonnet 5.5 takes 1M tokens, GPT-5.4 Mini 400K.
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 Sonnet 5.5 | Bars | GPT-5.4 Mini | Ties |
|---|---|---|---|---|
| Coding | 13 | 0 | ||
| Intelligence | 2 | 0 | ||
| Agentic work | 4 | 0 | ||
| Reasoning | 9 | 0 | ||
| Math | 6 | 0 | ||
| Knowledge | 5 | 0 | ||
| General | 21 | 0 | ||
| All shared boards | 60 | 0 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index130
| Benchmark | Claude Sonnet 5.5 | GPT-5.4 Mini | Leader |
|---|---|---|---|
| LiveBench coding | 94.7% | Not measured | Only Claude Sonnet 5.5 measured |
| LMArena WebDev | 1768 | 1397 | Claude Sonnet 5.5 |
| AA Coding Index | Not measured | 56.1 | Only GPT-5.4 Mini measured |
| LiveCodeBench (Vals AI) | Not measured | 81.5% | Only GPT-5.4 Mini measured |
| SWE-bench Verified (Vals AI) | Not measured | 73.0% | Only GPT-5.4 Mini measured |
| Terminal-Bench 4 (Vals AI) | 64.1% | 2.5% | Claude Sonnet 5.5 |
| Vibe Code Bench (Vals AI) | 92.4% | 48.0% | Claude Sonnet 5.5 |
| Code migration (Vals AI) | 69.8% | 12.9% | Claude Sonnet 5.5 |
| ProgramBench (Vals AI) | 6.5% | 0.0% | Claude Sonnet 5.5 |
| IOI (Vals AI) | 83.1% | Not measured | Only Claude Sonnet 5.5 measured |
| Terminal-Bench 2.1 (Vals AI) | 83.2% | 54.7% | Claude Sonnet 5.5 |
| SWE-bench Multimodal (Vals AI) | 58.3% | Not measured | Only Claude Sonnet 5.5 measured |
| CyberBench (Vals AI) | 59.6% | Not measured | Only Claude Sonnet 5.5 measured |
| SciCode | 61.0% | 52.1% | Claude Sonnet 5.5 |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 52.3% | Only GPT-5.4 Mini measured |
| FrontierCode | 52.1% | 27.0% | Claude Sonnet 5.5 |
| FrontierSWE | 61.9% | Not measured | Only Claude Sonnet 5.5 measured |
| WeirdML | Not measured | 60.3% | Only GPT-5.4 Mini measured |
| ALE-Bench | 1819.1 | 1188.6 | Claude Sonnet 5.5 |
| LMArena Text: coding | 1513 | 1438 | Claude Sonnet 5.5 |
| LMArena Text: software and IT | 1504 | 1436 | Claude Sonnet 5.5 |
| LMArena WebDev: React | 1817 | 1386 | Claude Sonnet 5.5 |
| LMArena WebDev: HTML | 1741 | 1419 | Claude Sonnet 5.5 |
| LMArena WebDev: from an image | 1737 | Not measured | Only Claude Sonnet 5.5 measured |
| SciCode | 13.8% | Not measured | Only Claude Sonnet 5.5 measured |
| Neither model has a result on the other 14 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Claude Sonnet 5.5 | GPT-5.4 Mini | Leader |
|---|---|---|---|
| AA Intelligence Index | 56.0 | 24.1 | Claude Sonnet 5.5 |
| Epoch Capabilities Index | 166.2 | 148.7 | Claude Sonnet 5.5 |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index40
| Benchmark | Claude Sonnet 5.5 | GPT-5.4 Mini | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 17.9 | Only GPT-5.4 Mini measured |
| Design Arena: agents building web apps | 1305 | Not measured | Only Claude Sonnet 5.5 measured |
| Design Arena: agents building mobile apps | 1260 | Not measured | Only Claude Sonnet 5.5 measured |
| Design Arena: spreadsheets | 1260 | Not measured | Only Claude Sonnet 5.5 measured |
| LiveBench research | 83.6% | Not measured | Only Claude Sonnet 5.5 measured |
| LiveBench browser | 84.0% | Not measured | Only Claude Sonnet 5.5 measured |
| LiveBench long agentic | 78.8% | Not measured | Only Claude Sonnet 5.5 measured |
| SREBench (Vals AI) | 30.1% | Not measured | Only Claude Sonnet 5.5 measured |
| Finance Agent (Vals AI) | 58.1% | 45.4% | Claude Sonnet 5.5 |
| Tax Agent Bench (Vals AI) | 42.0% | 8.8% | Claude Sonnet 5.5 |
| Legal Research Bench (Vals AI) | 48.1% | 12.5% | Claude Sonnet 5.5 |
| Harvey's Legal Agent Benchmark (Vals AI) | 10.4% | Not measured | Only Claude Sonnet 5.5 measured |
| Excel Modeling Benchmark (Vals AI) | 75.7% | 45.4% | Claude Sonnet 5.5 |
| BioMysteryBench (Vals AI) | 81.1% | Not measured | Only Claude Sonnet 5.5 measured |
| Terminal-Bench-Science (Vals AI) | 47.1% | Not measured | Only Claude Sonnet 5.5 measured |
| Vals Science Index | 59.1% | Not measured | Only Claude Sonnet 5.5 measured |
| τ²-Bench Telecom | Not measured | 83.3% | Only GPT-5.4 Mini measured |
| APEX-Agents | 75.5% | Not measured | Only Claude Sonnet 5.5 measured |
| DeepResearch Bench | Not measured | 36.3% | Only GPT-5.4 Mini measured |
| SpireClimb | 1.3% | Not measured | Only Claude Sonnet 5.5 measured |
| Vending-Bench 2 | $11,165 | Not measured | Only Claude Sonnet 5.5 measured |
| Neither model has a result on the other 20 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index90
| Benchmark | Claude Sonnet 5.5 | GPT-5.4 Mini | Leader |
|---|---|---|---|
| GPQA Diamond | Not measured | 83.1% | Only GPT-5.4 Mini measured |
| LiveBench reasoning | 91.6% | Not measured | Only Claude Sonnet 5.5 measured |
| MysteryMechanism (Vals AI) | 49.1% | Not measured | Only Claude Sonnet 5.5 measured |
| MMMU Pro (Vals AI) | Not measured | 79.3% | Only GPT-5.4 Mini measured |
| GPQA Diamond (Artificial Analysis run) | Not measured | 87.5% | Only GPT-5.4 Mini measured |
| Humanity's Last Exam (Artificial Analysis run) | 55.0% | 28.1% | Claude Sonnet 5.5 |
| GPQA Diamond (Epoch AI run) | 95.6% | 86.9% | Claude Sonnet 5.5 |
| ARC-AGI | Not measured | 63.7% | Only GPT-5.4 Mini measured |
| ARC-AGI-2 | Not measured | 18.9% | Only GPT-5.4 Mini measured |
| Chess puzzles | 59.0% | 23.0% | Claude Sonnet 5.5 |
| Mystery game puzzles | 65.0% | 11.0% | Claude Sonnet 5.5 |
| LMArena Text: hard prompts | 1494 | 1424 | Claude Sonnet 5.5 |
| LMArena Text: expert | 1531 | 1435 | Claude Sonnet 5.5 |
| LMArena Text: hard prompts in English | 1502 | 1420 | Claude Sonnet 5.5 |
| LMArena Vision: diagrams | 1329 | 1266 | Claude Sonnet 5.5 |
| LMArena Vision: homework | 1331 | 1285 | Claude Sonnet 5.5 |
| Kagi LLM Benchmark | Not measured | 37.9% | Only GPT-5.4 Mini 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 Sonnet 5.5 | GPT-5.4 Mini | Leader |
|---|---|---|---|
| AIME | Not measured | 95.6% | Only GPT-5.4 Mini measured |
| OTIS Mock AIME | 100.0% | 88.9% | Claude Sonnet 5.5 |
| FrontierMath, tiers 1 to 3 | 88.8% | 51.2% | Claude Sonnet 5.5 |
| FrontierMath, tier 4 | 80.5% | 9.8% | Claude Sonnet 5.5 |
| ProofBench | 100.0% | 21.0% | Claude Sonnet 5.5 |
| LMArena Text: math | 1497 | 1418 | Claude Sonnet 5.5 |
| LMArena Text: mathematical work | 1510 | 1429 | Claude Sonnet 5.5 |
| Neither model has a result on the other 6 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index50
| Benchmark | Claude Sonnet 5.5 | GPT-5.4 Mini | Leader |
|---|---|---|---|
| MMLU Pro | Not measured | 84.5% | Only GPT-5.4 Mini measured |
| CaseLaw (Vals AI) | Not measured | 51.7% | Only GPT-5.4 Mini measured |
| CorpFin (Vals AI) | Not measured | 60.9% | Only GPT-5.4 Mini measured |
| TaxEval (Vals AI) | Not measured | 71.2% | Only GPT-5.4 Mini measured |
| MortgageTax (Vals AI) | Not measured | 63.5% | Only GPT-5.4 Mini measured |
| MedCode (Vals AI) | 52.9% | Not measured | Only Claude Sonnet 5.5 measured |
| SimpleQA Verified | 46.5% | 29.4% | Claude Sonnet 5.5 |
| LMArena Text: science | 1483 | 1417 | Claude Sonnet 5.5 |
| LMArena Text: medicine and healthcare | 1464 | 1395 | Claude Sonnet 5.5 |
| LMArena Text: legal and government | 1483 | 1414 | Claude Sonnet 5.5 |
| LMArena Text: business and finance | 1468 | 1418 | Claude Sonnet 5.5 |
| Neither model has a result on the other 7 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
| Benchmark | Claude Sonnet 5.5 | GPT-5.4 Mini | Leader |
|---|---|---|---|
| Design Arena: websites | 1326 | Not measured | Only Claude Sonnet 5.5 measured |
| Design Arena: UI components | 1368 | Not measured | Only Claude Sonnet 5.5 measured |
| Design Arena: game dev | 1471 | Not measured | Only Claude Sonnet 5.5 measured |
| Design Arena: data viz | 1392 | Not measured | Only Claude Sonnet 5.5 measured |
| Design Arena: 3D | 1445 | Not measured | Only Claude Sonnet 5.5 measured |
| Design Arena: all code categories | 1364 | Not measured | Only Claude Sonnet 5.5 measured |
| LiveBench design | 85.0% | Not measured | Only Claude Sonnet 5.5 measured |
| Neither model has a result on the other 2 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index210
| Benchmark | Claude Sonnet 5.5 | GPT-5.4 Mini | Leader |
|---|---|---|---|
| Vals Index | 70.8% | Not measured | Only Claude Sonnet 5.5 measured |
| LiveBench data analysis | 59.5% | Not measured | Only Claude Sonnet 5.5 measured |
| Public Benefits Bench (Vals AI) | 67.2% | Not measured | Only Claude Sonnet 5.5 measured |
| SAGE (Vals AI) | 51.8% | 50.8% | Claude Sonnet 5.5 |
| MedScribe (Vals AI) | 91.1% | Not measured | Only Claude Sonnet 5.5 measured |
| Vals Multimodal Index | Not measured | 54.2% | Only GPT-5.4 Mini measured |
| IFBench | Not measured | 73.3% | Only GPT-5.4 Mini measured |
| Long context reasoning (AA-LCR) | 82.7% | 77.0% | Claude Sonnet 5.5 |
| DTBench | 97.3% | 80.0% | Claude Sonnet 5.5 |
| LMCA | 68.3% | 40.8% | Claude Sonnet 5.5 |
| Furniture assembly | 75.0% | Not measured | Only Claude Sonnet 5.5 measured |
| LMArena Text | 1473 | 1412 | Claude Sonnet 5.5 |
| LMArena Text: instruction following | 1492 | 1405 | Claude Sonnet 5.5 |
| LMArena Text: creative writing | 1467 | 1370 | Claude Sonnet 5.5 |
| LMArena Vision | 1287 | 1245 | Claude Sonnet 5.5 |
| LMArena Text: English | 1484 | 1412 | Claude Sonnet 5.5 |
| LMArena Text: non-English | 1456 | 1405 | Claude Sonnet 5.5 |
| LMArena Text: Chinese | 1535 | 1446 | Claude Sonnet 5.5 |
| LMArena Text: German | Not measured | 1408 | Only GPT-5.4 Mini measured |
| LMArena Text: Japanese | Not measured | 1375 | Only GPT-5.4 Mini measured |
| LMArena Text: Korean | Not measured | 1370 | Only GPT-5.4 Mini measured |
| LMArena Text: French | 1499 | 1441 | Claude Sonnet 5.5 |
| LMArena Text: Spanish | 1451 | 1405 | Claude Sonnet 5.5 |
| LMArena Text: Russian | 1454 | 1417 | Claude Sonnet 5.5 |
| LMArena Text: Polish | Not measured | 1415 | Only GPT-5.4 Mini measured |
| LMArena Text: multi-turn | 1465 | 1429 | Claude Sonnet 5.5 |
| LMArena Text: longer queries | 1495 | 1408 | Claude Sonnet 5.5 |
| LMArena Text: writing and language | 1474 | 1393 | Claude Sonnet 5.5 |
| LMArena Text: entertainment and media | 1461 | 1373 | Claude Sonnet 5.5 |
| LMArena Vision: English | 1289 | 1244 | Claude Sonnet 5.5 |
| LMArena Vision: Chinese | Not measured | 1286 | Only GPT-5.4 Mini measured |
| LMArena Vision: reading text | 1297 | 1253 | Claude Sonnet 5.5 |
| LMArena Vision: humor | Not measured | 1235 | Only GPT-5.4 Mini measured |
| LMArena Vision: recognizing things | Not measured | 1218 | Only GPT-5.4 Mini measured |
| LMArena Vision: creative writing | 1242 | 1222 | Claude Sonnet 5.5 |
| 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 | Claude Sonnet 5.5 | GPT-5.4 Mini | 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
GPT-5.4 Mini costs less
Repository review
50K input + 3K output tokens
GPT-5.4 Mini costs less
Agent loop
200K input + 10K output tokens
GPT-5.4 Mini 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 Sonnet 5.5 | OpenAI GPT-5.4 Mini |
|---|---|---|
| Maker | Anthropic | OpenAI |
| Weights | No public weights listed | No public weights listed |
| Released | September 28, 2026 | March 17, 2026 |
| Context window | 1M | 400K |
| Input $/Mtok | $2.00 | $0.75 |
| Output $/Mtok | $10.00 | $4.50 |
| Coding index | 87.3 | 35.8 |
| Independent measurements behind it | 5 | 5 |
| Cost per solved task | Not computable | $0.70 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Claude Sonnet 5.5 better than GPT-5.4 Mini for coding?
Pick Claude Sonnet 5.5 for coding: it scores 87.3 on the Vibe Coding Index against 35.8 for GPT-5.4 Mini. Pick GPT-5.4 Mini for output tokens 2.2x cheaper ($4.50 vs $10.00 per million).
Which is cheaper, Claude Sonnet 5.5 or GPT-5.4 Mini?
Claude Sonnet 5.5 costs $2.00 per million input tokens and $10.00 per million output tokens; GPT-5.4 Mini costs $0.75 and $4.50. Cost per solved task is known for GPT-5.4 Mini only.
Which has the bigger context window, Claude Sonnet 5.5 or GPT-5.4 Mini?
Claude Sonnet 5.5, with 1M tokens against 400K.
How many benchmarks is the Claude Sonnet 5.5 vs GPT-5.4 Mini comparison based on?
60 benchmarks have a result for both models as of October 22, 2026. Claude Sonnet 5.5's coding index rests on 5 independent benchmarks and GPT-5.4 Mini's on 5. Where only one model has a result, the page says so instead of naming a winner.