Model comparison
Claude Haiku 4.5 vs o3
o3 is the stronger model overall, scoring 47.5 to 39.5 on the Noometry Index. Claude Haiku 4.5 costs 1.8× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Last verified . 41 shared benchmarks.
Summary
- They share 41 benchmarks with published results for both. Claude Haiku 4.5 scores higher in 0 categories and o3 in 10 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where o3 leads 32.0 to 15.1.
- The biggest single-benchmark swing is SimpleQA Verified: 13.2% for Claude Haiku 4.5 and 49.4% for o3.
- Claude Haiku 4.5 is cheaper at $1 / $5 per million input/output tokens, against $2 / $8 for o3.
Side by side
| Claude Haiku 4.5 | o3 | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 39.5 | 47.5 |
| Released | 2025-10-15 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | 200K |
| Max output | 64K | 100K |
| Input $ / M tokens | $1 | $2 |
| Output $ / M tokens | $5 | $8 |
| Results tracked | 53 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3 leads
Claude Haiku 4.5: 44.0 (#78), o3: 46.8 (#64)
| Benchmark | Claude Haiku 4.5 | o3 |
|---|---|---|
| SWE-bench Verified (bash only) | 66.6% | 58.4% |
| WeirdML | 45.4% | 52.4% |
| LMArena Coding | 1453 | 1408 |
| ALE-Bench | 653.48 | 933.55 |
| SWE-bench Verified | — | 62.3% |
| Aider Polyglot | — | 81.3% |
| LMArena WebDev | 1330 | — |
| SWE-bench Multilingual | 64.7% | — |
| SciCode | 43.3% | — |
| GSO | — | 8.8% |
| CadEval | — | 74% |
Agentic & Tool Use Too close to call
Claude Haiku 4.5: 33.6 (#52), o3: 34.5 (#44)
| Benchmark | Claude Haiku 4.5 | o3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 68.7% | 63% |
| DeepResearch Bench | 45.5% | 45.2% |
| Terminal-Bench | 35.5% | — |
| GDPval | — | 30.8% |
| OSWorld | — | 23% |
| BALROG | 31.2% | — |
| ExploitBench | 13.7% | — |
| LMArena Search | — | 1144 |
| METR Time Horizons | — | 65.4% |
| Vending-Bench 2 | 458.89 | — |
Reasoning o3 leads
Claude Haiku 4.5: 15.1 (#320), o3: 32.0 (#78)
| Benchmark | Claude Haiku 4.5 | o3 |
|---|---|---|
| ARC-AGI-2 | 4% | 6.5% |
| ARC-AGI-1 | 47.7% | 60.8% |
| CritPt | 0% | 1.4% |
| Chess Puzzles | 8% | 38% |
| LMArena Hard Prompts | 1420 | 1402 |
| DTBench | 73.6% | 84.8% |
| LMCA | 30.9% | 39.7% |
| Epoch Capabilities Index | 142.41 | 146.86 |
| ForecastBench | 61.4 | 62.5 |
| SimpleBench | — | 53.1% |
| Kagi LLM Benchmark | — | 67.6% |
| NYT Connections (extended) | 14.3% | — |
| EnigmaEval | — | 13.1% |
| Mystery Game Puzzles | — | 29% |
Math o3 leads
Claude Haiku 4.5: 44.9 (#78), o3: 50.2 (#58)
| Benchmark | Claude Haiku 4.5 | o3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.7% | 84.4% |
| Omni-MATH | 56.1% | 71.4% |
| LMArena Math | 1396 | 1426 |
| MATH Level 5 | 96.4% | 97.8% |
| FrontierMath (Feb 2025 set) | 5.9% | 18.7% |
| FrontierMath Tier 4 (v1) | 2.1% | 2.1% |
| FrontierMath (Tiers 1-3) | — | 33.3% |
Knowledge o3 leads
Claude Haiku 4.5: 37.7 (#153), o3: 54.6 (#52)
| Benchmark | Claude Haiku 4.5 | o3 |
|---|---|---|
| GPQA Diamond | 71.2% | 81.8% |
| SimpleQA Verified | 13.2% | 49.4% |
| MMLU-Pro | 77.7% | 85.9% |
| GPQA (HELM) | 60.5% | 75.3% |
| LMArena Expert | 1442 | 1402 |
| Humanity's Last Exam | — | 20.3% |
| Confabulations | — | 14.4% |
| Vectara Hallucination Rate | 9.8% | — |
Multimodal o3 leads
Claude Haiku 4.5: 26.8 (#118), o3: 41.4 (#36)
| Benchmark | Claude Haiku 4.5 | o3 |
|---|---|---|
| LMArena Vision | — | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
| Blueprint-Bench 2 | 0% | — |
| LMArena Document | 1420 | — |
Multilingual o3 leads
Claude Haiku 4.5: 49.9 (#129), o3: 51.7 (#105)
| Benchmark | Claude Haiku 4.5 | o3 |
|---|---|---|
| LMArena Non-English | 1377 | 1401 |
| LMArena Chinese | 1417 | 1437 |
| LMArena French | 1408 | 1430 |
| LMArena German | 1375 | 1420 |
| LMArena Japanese | 1339 | 1403 |
| LMArena Korean | 1347 | 1370 |
| LMArena Russian | 1381 | 1406 |
| LMArena Spanish | 1420 | 1395 |
Instruction Following o3 leads
Claude Haiku 4.5: 71.4 (#149), o3: 72.8 (#127)
| Benchmark | Claude Haiku 4.5 | o3 |
|---|---|---|
| IFEval | 80.1% | 86.9% |
| LMArena Instruction Following | 1414 | 1368 |
Long Context o3 leads
Claude Haiku 4.5: 43.6 (#92), o3: 53.3 (#6)
| Benchmark | Claude Haiku 4.5 | o3 |
|---|---|---|
| LMArena Longer Query | 1427 | 1372 |
| Fiction.LiveBench | — | 88.9% |
| CL-bench | — | 17.8% |
Writing & Preference o3 leads
Claude Haiku 4.5: 57.9 (#123), o3: 63.5 (#64)
| Benchmark | Claude Haiku 4.5 | o3 |
|---|---|---|
| LMArena Text | 1396 | 1410 |
| LMArena Creative Writing | 1372 | 1359 |
| WildBench | 83.9% | 86.1% |
| LMArena Multi-Turn | 1409 | 1405 |
| Short-Story Creative Writing | — | 83.9% |
| EQ-Bench Creative Writing | — | 1676 |
| EQ-Bench 4 | 1064 | — |
Frequently asked questions
Is Claude Haiku 4.5 better than o3?
o3 is the stronger model overall, scoring 47.5 to 39.5 on the Noometry Index. Claude Haiku 4.5 costs 1.8× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Which is cheaper, Claude Haiku 4.5 or o3?
Claude Haiku 4.5 is cheaper. It lists at $1 per million input tokens and $5 per million output tokens; o3 lists at $2 and $8.
Is Claude Haiku 4.5 or o3 better for coding?
o3 scores higher on coding benchmarks: 46.8 versus 44.0 in the Noometry coding category.
Which has the bigger context window?
Both accept 200K tokens.
How many benchmarks do Claude Haiku 4.5 and o3 share?
41 benchmarks have published results for both models. Claude Haiku 4.5 has 53 scored results on Noometry and o3 has 63.