Claude Haiku 5.5 vs GPT-6 Sol
Updated October 21, 2026. 57 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.
Decision reading
Pick GPT-6 Sol for coding: it scores 82.8 on the Vibe Coding Index against 75.9 for Claude Haiku 5.5.
Pick Claude Haiku 5.5 when output tokens cost 20.0x less, or when what matters is Vibe Code Bench (Vals AI) or ProgramBench (Vals AI), where it leads.
- Shared results
- 57
- Claude Haiku 5.5 only
- 5
- GPT-6 Sol only
- 43
- Like-for-like categories
- 8 of 9
Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.
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, four public coding benchmarks
GPT-6 Sol
GPT-6 Sol scores 82.8 against 75.9.
2 benchmarks behind each score
Intelligence
AA Intelligence Index
GPT-6 Sol
GPT-6 Sol leads, 47.6 against 43.4.
Same board, both measured
Web design
Design Arena: websites
GPT-6 Sol
GPT-6 Sol leads, 1292 against 1263.
Same board, both measured
General
Vals Index
No call
Only Claude Haiku 5.5 has a Vals Index result, 58.7%.
Directional only
Long documents
Listed context window
GPT-6 Sol
GPT-6 Sol takes 1.05M tokens against 1M.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Claude Haiku 5.5
Claude Haiku 5.5 $0.0003, GPT-6 Sol $0.0070 per turn.
List prices
At a glance
- Coding index
- Claude Haiku 5.5 75.9 from 2 benchmarks, GPT-6 Sol 82.8 from 2 benchmarks.
- Widest category
- Knowledge, 20.7 points apart on average over 2 shared benchmarks.
- Widest single gap
- Vending-Bench 2: Claude Haiku 5.5 $3,602 against GPT-6 Sol $14,428, 67.7 points apart on a 0 to 100 scale.
- Output price
- Claude Haiku 5.5 $0.50 against GPT-6 Sol $10.00 per million output tokens, 20.0x apart.
- Context window
- Claude Haiku 5.5 takes 1M tokens, GPT-6 Sol 1.05M.
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 Haiku 5.5 | Bars | GPT-6 Sol | Ties |
|---|---|---|---|---|
| Coding | 2 | 10 | ||
| Intelligence | 0 | 2 | ||
| Agentic work | 5 | 8 | ||
| Reasoning | 1 | 8 | ||
| Math | 1 | 3 | ||
| Knowledge | 1 | 1 | ||
| Web design | 5 | 2 | ||
| General | 3 | 5 | ||
| All shared boards | 18 | 39 |
CodingLike-for-likeFour of these boards feed the Vibe Coding Index210
| Benchmark | Claude Haiku 5.5 | GPT-6 Sol | Leader |
|---|---|---|---|
| LiveBench coding | 91.4% | 93.1% | GPT-6 Sol |
| LMArena WebDev | 1586 | 1683 | GPT-6 Sol |
| Terminal-Bench 4 (Vals AI) | 36.4% | 44.4% | GPT-6 Sol |
| Vibe Code Bench (Vals AI) | 90.4% | 87.8% | Claude Haiku 5.5 |
| Code migration (Vals AI) | 44.6% | 57.2% | GPT-6 Sol |
| ProgramBench (Vals AI) | 2.5% | 2.0% | Claude Haiku 5.5 |
| IOI (Vals AI) | 47.3% | 82.6% | GPT-6 Sol |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 83.2% | Only GPT-6 Sol measured |
| SWE-bench Multimodal (Vals AI) | 44.0% | Not measured | Only Claude Haiku 5.5 measured |
| Vibe Code Bench 1-100 (Vals AI) | 20.4% | Not measured | Only Claude Haiku 5.5 measured |
| CyberBench (Vals AI) | 75.5% | 78.0% | GPT-6 Sol |
| SciCode | 55.0% | 57.6% | GPT-6 Sol |
| FrontierCode | 46.4% | 49.3% | GPT-6 Sol |
| FrontierSWE | 43.8% | Not measured | Only Claude Haiku 5.5 measured |
| ALE-Bench | Not measured | 2462.0 | Only GPT-6 Sol measured |
| LMArena Text: coding | Not measured | 1448 | Only GPT-6 Sol measured |
| LMArena Text: software and IT | Not measured | 1433 | Only GPT-6 Sol measured |
| LMArena WebDev: React | 1605 | 1685 | GPT-6 Sol |
| LMArena WebDev: HTML | 1569 | 1685 | GPT-6 Sol |
| LMArena WebDev: from an image | Not measured | 1660 | Only GPT-6 Sol measured |
| Neither model has a result on the other 18 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | Claude Haiku 5.5 | GPT-6 Sol | Leader |
|---|---|---|---|
| AA Intelligence Index | 43.4 | 47.6 | GPT-6 Sol |
| Epoch Capabilities Index | 159.3 | 163.7 | GPT-6 Sol |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index58
| Benchmark | Claude Haiku 5.5 | GPT-6 Sol | Leader |
|---|---|---|---|
| Design Arena: agents building web apps | Not measured | 1232 | Only GPT-6 Sol measured |
| Design Arena: agents building full-stack apps | Not measured | 1238 | Only GPT-6 Sol measured |
| Design Arena: agents building mobile apps | Not measured | 1216 | Only GPT-6 Sol measured |
| LiveBench research | 81.9% | 69.3% | Claude Haiku 5.5 |
| LiveBench browser | 80.0% | 92.0% | GPT-6 Sol |
| LiveBench long agentic | 57.6% | 66.9% | GPT-6 Sol |
| SREBench (Vals AI) | 5.7% | 30.5% | GPT-6 Sol |
| Finance Agent (Vals AI) | 54.1% | 49.0% | Claude Haiku 5.5 |
| Tax Agent Bench (Vals AI) | 31.9% | 15.0% | Claude Haiku 5.5 |
| Legal Research Bench (Vals AI) | 43.3% | 28.9% | Claude Haiku 5.5 |
| Harvey's Legal Agent Benchmark (Vals AI) | 4.2% | Not measured | Only Claude Haiku 5.5 measured |
| Excel Modeling Benchmark (Vals AI) | 59.0% | 71.5% | GPT-6 Sol |
| BioMysteryBench (Vals AI) | 70.4% | 74.8% | GPT-6 Sol |
| Terminal-Bench-Science (Vals AI) | 10.0% | 30.0% | GPT-6 Sol |
| Vals Science Index | 38.5% | 45.0% | GPT-6 Sol |
| APEX-Agents | 56.1% | 54.3% | Claude Haiku 5.5 |
| SpireClimb | Not measured | 7.5% | Only GPT-6 Sol measured |
| Vending-Bench 2 | $3,602 | $14,428 | GPT-6 Sol |
| Neither model has a result on the other 21 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index18
| Benchmark | Claude Haiku 5.5 | GPT-6 Sol | Leader |
|---|---|---|---|
| LiveBench reasoning | 82.4% | 88.7% | GPT-6 Sol |
| MysteryMechanism (Vals AI) | 35.1% | 30.2% | Claude Haiku 5.5 |
| Humanity's Last Exam (Artificial Analysis run) | 44.4% | 47.9% | GPT-6 Sol |
| GPQA Diamond (Epoch AI run) | 89.6% | 94.3% | GPT-6 Sol |
| ARC-AGI | 89.0% | 95.5% | GPT-6 Sol |
| ARC-AGI-2 | 61.3% | 89.6% | GPT-6 Sol |
| Chess puzzles | 55.0% | 56.0% | GPT-6 Sol |
| Mystery game puzzles | 30.0% | 56.0% | GPT-6 Sol |
| EBR-bench | 43.8% | 53.3% | GPT-6 Sol |
| LMArena Text: hard prompts | Not measured | 1422 | Only GPT-6 Sol measured |
| LMArena Text: expert | Not measured | 1437 | Only GPT-6 Sol measured |
| LMArena Text: hard prompts in English | Not measured | 1420 | Only GPT-6 Sol measured |
| LMArena Vision: diagrams | Not measured | 1270 | Only GPT-6 Sol measured |
| LMArena Vision: homework | Not measured | 1286 | Only GPT-6 Sol measured |
| Neither model has a result on the other 7 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index13
| Benchmark | Claude Haiku 5.5 | GPT-6 Sol | Leader |
|---|---|---|---|
| OTIS Mock AIME | 98.9% | 100.0% | GPT-6 Sol |
| FrontierMath, tiers 1 to 3 | 75.1% | 89.8% | GPT-6 Sol |
| FrontierMath, tier 4 | 46.3% | 90.0% | GPT-6 Sol |
| ProofBench | 89.0% | 83.0% | Claude Haiku 5.5 |
| LMArena Text: math | Not measured | 1407 | Only GPT-6 Sol measured |
| LMArena Text: mathematical work | Not measured | 1424 | Only GPT-6 Sol measured |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index11
| Benchmark | Claude Haiku 5.5 | GPT-6 Sol | Leader |
|---|---|---|---|
| MedCode (Vals AI) | 51.5% | 47.1% | Claude Haiku 5.5 |
| SimpleQA Verified | 23.8% | 60.7% | GPT-6 Sol |
| LMArena Text: science | Not measured | 1389 | Only GPT-6 Sol measured |
| LMArena Text: medicine and healthcare | Not measured | 1366 | Only GPT-6 Sol measured |
| LMArena Text: legal and government | Not measured | 1396 | Only GPT-6 Sol measured |
| LMArena Text: business and finance | Not measured | 1380 | Only GPT-6 Sol measured |
| Neither model has a result on the other 11 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index52
| Benchmark | Claude Haiku 5.5 | GPT-6 Sol | Leader |
|---|---|---|---|
| Design Arena: websites | 1263 | 1292 | GPT-6 Sol |
| Design Arena: UI components | 1295 | 1292 | Claude Haiku 5.5 |
| Design Arena: game dev | 1315 | 1313 | Claude Haiku 5.5 |
| Design Arena: data viz | 1292 | 1269 | Claude Haiku 5.5 |
| Design Arena: 3D | 1343 | 1322 | Claude Haiku 5.5 |
| Design Arena: all code categories | 1279 | 1301 | GPT-6 Sol |
| Design Arena: SVG | Not measured | 1318 | Only GPT-6 Sol measured |
| Design Arena: ASCII art | Not measured | 1226 | Only GPT-6 Sol measured |
| LiveBench design | 83.1% | 82.2% | Claude Haiku 5.5 |
GeneralLike-for-likePublic boards, shown as published, kept out of the index35
| Benchmark | Claude Haiku 5.5 | GPT-6 Sol | Leader |
|---|---|---|---|
| Vals Index | 58.7% | Not measured | Only Claude Haiku 5.5 measured |
| LiveBench data analysis | 53.1% | 81.2% | GPT-6 Sol |
| Public Benefits Bench (Vals AI) | 63.2% | 56.6% | Claude Haiku 5.5 |
| SAGE (Vals AI) | 45.8% | 44.8% | Claude Haiku 5.5 |
| MedScribe (Vals AI) | 89.5% | 82.0% | Claude Haiku 5.5 |
| Long context reasoning (AA-LCR) | 82.7% | 83.7% | GPT-6 Sol |
| DTBench | 95.0% | 97.3% | GPT-6 Sol |
| LMCA | 51.1% | 59.1% | GPT-6 Sol |
| Furniture assembly | 47.5% | 58.3% | GPT-6 Sol |
| LMArena Text | Not measured | 1398 | Only GPT-6 Sol measured |
| LMArena Text: instruction following | Not measured | 1413 | Only GPT-6 Sol measured |
| LMArena Text: creative writing | Not measured | 1378 | Only GPT-6 Sol measured |
| LMArena Vision | Not measured | 1240 | Only GPT-6 Sol measured |
| LMArena Text: English | Not measured | 1396 | Only GPT-6 Sol measured |
| LMArena Text: non-English | Not measured | 1388 | Only GPT-6 Sol measured |
| LMArena Text: Chinese | Not measured | 1412 | Only GPT-6 Sol measured |
| LMArena Text: German | Not measured | 1379 | Only GPT-6 Sol measured |
| LMArena Text: Japanese | Not measured | 1377 | Only GPT-6 Sol measured |
| LMArena Text: Korean | Not measured | 1336 | Only GPT-6 Sol measured |
| LMArena Text: French | Not measured | 1399 | Only GPT-6 Sol measured |
| LMArena Text: Spanish | Not measured | 1387 | Only GPT-6 Sol measured |
| LMArena Text: Russian | Not measured | 1402 | Only GPT-6 Sol measured |
| LMArena Text: Polish | Not measured | 1372 | Only GPT-6 Sol measured |
| LMArena Text: multi-turn | Not measured | 1414 | Only GPT-6 Sol measured |
| LMArena Text: longer queries | Not measured | 1416 | Only GPT-6 Sol measured |
| LMArena Text: writing and language | Not measured | 1389 | Only GPT-6 Sol measured |
| LMArena Text: entertainment and media | Not measured | 1373 | Only GPT-6 Sol measured |
| LMArena Vision: English | Not measured | 1229 | Only GPT-6 Sol measured |
| LMArena Vision: reading text | Not measured | 1258 | Only GPT-6 Sol measured |
| LMArena Vision: creative writing | Not measured | 1219 | Only GPT-6 Sol measured |
| Neither model has a result on the other 16 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Claude Haiku 5.5 | GPT-6 Sol | 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 Haiku 5.5 costs less
Repository review
50K input + 3K output tokens
Claude Haiku 5.5 costs less
Agent loop
200K input + 10K output tokens
Claude Haiku 5.5 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 Haiku 5.5 | OpenAI GPT-6 Sol |
|---|---|---|
| Maker | Anthropic | OpenAI |
| Weights | No public weights listed | No public weights listed |
| Released | October 7, 2026 | September 22, 2026 |
| Context window | 1M | 1.05M |
| Input $/Mtok | $0.10 | $2.00 |
| Output $/Mtok | $0.50 | $10.00 |
| Coding index | 75.9 | 82.8 |
| Independent measurements behind it | 2 | 2 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Claude Haiku 5.5 better than GPT-6 Sol for coding?
Pick GPT-6 Sol for coding: it scores 82.8 on the Vibe Coding Index against 75.9 for Claude Haiku 5.5. Pick Claude Haiku 5.5 when output tokens cost 20.0x less, or when what matters is Vibe Code Bench (Vals AI) or ProgramBench (Vals AI), where it leads.
Which is cheaper, Claude Haiku 5.5 or GPT-6 Sol?
Claude Haiku 5.5 costs $0.10 per million input tokens and $0.50 per million output tokens; GPT-6 Sol costs $2.00 and $10.00. Cost per solved task is known for neither of them.
Which has the bigger context window, Claude Haiku 5.5 or GPT-6 Sol?
GPT-6 Sol, with 1.05M tokens against 1M.
How many benchmarks is the Claude Haiku 5.5 vs GPT-6 Sol comparison based on?
57 benchmarks have a result for both models as of October 21, 2026. Claude Haiku 5.5's coding index rests on 2 independent benchmarks and GPT-6 Sol's on 2. Where only one model has a result, the page says so instead of naming a winner.