Model comparison
Claude Haiku 4.5 vs GPT-5.1
GPT-5.1 is the stronger model overall, scoring 48.8 to 39.3 on the Noometry Index. Claude Haiku 4.5 costs 1.7× less per token, which makes it the better buy when GPT-5.1's lead doesn't matter for your workload.
Last verified . 45 shared benchmarks.
Summary
- They share 45 benchmarks with published results for both. Claude Haiku 4.5 scores higher in 1 category and GPT-5.1 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.1 leads 39.9 to 15.1.
- The biggest single-benchmark swing is SimpleQA Verified: 13.2% for Claude Haiku 4.5 and 48% for GPT-5.1.
- Claude Haiku 4.5 is cheaper at $1 / $5 per million input/output tokens, against $1.25 / $10 for GPT-5.1.
- GPT-5.1 accepts more context: 400K tokens versus 200K.
Side by side
| Claude Haiku 4.5 | GPT-5.1 | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 39.3 | 48.8 |
| Released | 2025-10-15 | 2025-11-13 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | 400K |
| Max output | 64K | 128K |
| Input $ / M tokens | $1 | $1.25 |
| Output $ / M tokens | $5 | $10 |
| Results tracked | 53 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.1 leads
Claude Haiku 4.5: 43.9 (#71), GPT-5.1: 46.9 (#59)
| Benchmark | Claude Haiku 4.5 | GPT-5.1 |
|---|---|---|
| SWE-bench Verified (bash only) | 66.6% | 66% |
| LMArena WebDev | 1330 | 1395 |
| SciCode | 43.3% | 43.3% |
| WeirdML | 45.4% | 60.8% |
| LMArena Coding | 1454 | 1454 |
| ALE-Bench | 653.48 | 1,192 |
| SWE-bench Verified | — | 68% |
| SWE-bench Multilingual | 64.7% | — |
| GSO | — | 13.7% |
| LiveBench Coding | — | 72.5% |
Agentic & Tool Use Too close to call
Claude Haiku 4.5: 33.5 (#50), GPT-5.1: 32.5 (#55)
| Benchmark | Claude Haiku 4.5 | GPT-5.1 |
|---|---|---|
| Terminal-Bench | 35.5% | 47.6% |
| DeepResearch Bench | 45.5% | 42.8% |
| Vending-Bench 2 | 458.89 | 1,473 |
| Berkeley Function Calling Leaderboard | 68.7% | — |
| BALROG | 31.2% | — |
| ExploitBench | 13.7% | — |
| LMArena Search | — | 1199 |
Reasoning GPT-5.1 leads
Claude Haiku 4.5: 15.1 (#310), GPT-5.1: 39.9 (#50)
| Benchmark | Claude Haiku 4.5 | GPT-5.1 |
|---|---|---|
| ARC-AGI-2 | 4% | 17.6% |
| ARC-AGI-1 | 47.7% | 72.8% |
| CritPt | 0% | 4.9% |
| Chess Puzzles | 8% | 32% |
| LMArena Hard Prompts | 1420 | 1457 |
| DTBench | 73.6% | 90.1% |
| LMCA | 30.9% | 43.9% |
| Epoch Capabilities Index | 142.42 | 149.66 |
| ForecastBench | 61.4 | 58.1 |
| SimpleBench | — | 53.2% |
| NYT Connections (extended) | 14.3% | — |
| EnigmaEval | — | 11.2% |
| LiveBench Reasoning | — | 95.8% |
| Mystery Game Puzzles | — | 19% |
| LiveBench Data Analysis | — | 72.1% |
| LiveBench | — | 78.8% |
Math GPT-5.1 leads
Claude Haiku 4.5: 43.0 (#69), GPT-5.1: 50.3 (#45)
| Benchmark | Claude Haiku 4.5 | GPT-5.1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.7% | 88.6% |
| Omni-MATH | 56.1% | 46.4% |
| LMArena Math | 1395 | 1446 |
| FrontierMath (Feb 2025 set) | 5.9% | 31% |
| FrontierMath Tier 4 (v1) | 2.1% | 12.5% |
| LiveBench Math | — | 94.5% |
| MATH Level 5 | 96.4% | — |
Knowledge GPT-5.1 leads
Claude Haiku 4.5: 37.6 (#146), GPT-5.1: 50.5 (#65)
| Benchmark | Claude Haiku 4.5 | GPT-5.1 |
|---|---|---|
| GPQA Diamond | 71.2% | 87.6% |
| SimpleQA Verified | 13.2% | 48% |
| MMLU-Pro | 77.7% | 57.9% |
| Vectara Hallucination Rate | 9.8% | 10.9% |
| GPQA (HELM) | 60.5% | 44.2% |
| LMArena Expert | 1443 | 1470 |
| Humanity's Last Exam | — | 23.7% |
Multimodal GPT-5.1 leads
Claude Haiku 4.5: 27.0 (#111), GPT-5.1: 44.0 (#17)
| Benchmark | Claude Haiku 4.5 | GPT-5.1 |
|---|---|---|
| LMArena Document | 1420 | 1403 |
| LMArena Vision | — | 1250 |
| VPCT | — | 58.7% |
| Blueprint-Bench 2 | 0% | — |
Multilingual GPT-5.1 leads
Claude Haiku 4.5: 50.1 (#124), GPT-5.1: 54.0 (#51)
| Benchmark | Claude Haiku 4.5 | GPT-5.1 |
|---|---|---|
| LMArena Non-English | 1376 | 1431 |
| LMArena Chinese | 1415 | 1495 |
| LMArena French | 1408 | 1451 |
| LMArena German | 1377 | 1439 |
| LMArena Japanese | 1339 | 1453 |
| LMArena Korean | 1346 | 1401 |
| LMArena Russian | 1380 | 1435 |
| LMArena Spanish | 1420 | 1430 |
Instruction Following GPT-5.1 leads
Claude Haiku 4.5: 71.6 (#144), GPT-5.1: 84.0 (#1)
| Benchmark | Claude Haiku 4.5 | GPT-5.1 |
|---|---|---|
| IFEval | 80.1% | 93.5% |
| LMArena Instruction Following | 1414 | 1442 |
| LiveBench Instruction Following | — | 93.3% |
Long Context GPT-5.1 leads
Claude Haiku 4.5: 43.6 (#86), GPT-5.1: 47.6 (#14)
| Benchmark | Claude Haiku 4.5 | GPT-5.1 |
|---|---|---|
| LMArena Longer Query | 1427 | 1448 |
| CL-bench | — | 23.7% |
| CL-bench Life | — | 17.3% |
Writing & Preference GPT-5.1 leads
Claude Haiku 4.5: 58.1 (#119), GPT-5.1: 64.7 (#53)
| Benchmark | Claude Haiku 4.5 | GPT-5.1 |
|---|---|---|
| LMArena Text | 1396 | 1443 |
| LMArena Creative Writing | 1371 | 1427 |
| WildBench | 83.9% | 86.3% |
| LMArena Multi-Turn | 1409 | 1449 |
| EQ-Bench 4 | 1064 | — |
| LiveBench Language | — | 80.2% |
Frequently asked questions
Is Claude Haiku 4.5 better than GPT-5.1?
GPT-5.1 is the stronger model overall, scoring 48.8 to 39.3 on the Noometry Index. Claude Haiku 4.5 costs 1.7× less per token, which makes it the better buy when GPT-5.1's lead doesn't matter for your workload.
Which is cheaper, Claude Haiku 4.5 or GPT-5.1?
Claude Haiku 4.5 is cheaper. It lists at $1 per million input tokens and $5 per million output tokens; GPT-5.1 lists at $1.25 and $10.
Is Claude Haiku 4.5 or GPT-5.1 better for coding?
GPT-5.1 scores higher on coding benchmarks: 46.9 versus 43.9 in the Noometry coding category.
Which has the bigger context window?
GPT-5.1 does, with 400K tokens against 200K.
How many benchmarks do Claude Haiku 4.5 and GPT-5.1 share?
45 benchmarks have published results for both models. Claude Haiku 4.5 has 53 scored results on Noometry and GPT-5.1 has 63.