Model comparison
Claude Haiku 4.5 vs GPT-5 Mini
GPT-5 Mini is the stronger model overall, scoring 41.7 to 39.3 on the Noometry Index.
Last verified . 45 shared benchmarks.
Summary
- They share 45 benchmarks with published results for both. Claude Haiku 4.5 scores higher in 5 categories and GPT-5 Mini in 5 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5 Mini leads 24.1 to 15.1.
- The biggest single-benchmark swing is SWE-bench Multilingual: 64.7% for Claude Haiku 4.5 and 39.7% for GPT-5 Mini.
- GPT-5 Mini is cheaper at $0.25 / $2 per million input/output tokens, against $1 / $5 for Claude Haiku 4.5.
- GPT-5 Mini accepts more context: 400K tokens versus 200K.
Side by side
| Claude Haiku 4.5 | GPT-5 Mini | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 39.3 | 41.7 |
| Released | 2025-10-15 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | 400K |
| Max output | 64K | 128K |
| Input $ / M tokens | $1 | $0.25 |
| Output $ / M tokens | $5 | $2 |
| Results tracked | 53 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Haiku 4.5 leads
Claude Haiku 4.5: 43.9 (#71), GPT-5 Mini: 40.0 (#134)
| Benchmark | Claude Haiku 4.5 | GPT-5 Mini |
|---|---|---|
| SWE-bench Verified (bash only) | 66.6% | 59.8% |
| SWE-bench Multilingual | 64.7% | 39.7% |
| SciCode | 43.3% | 39.2% |
| WeirdML | 45.4% | 52.7% |
| LMArena Coding | 1454 | 1406 |
| ALE-Bench | 653.48 | 799.77 |
| SWE-bench Verified | — | 64.7% |
| LMArena WebDev | 1330 | — |
| AlgoTune | — | 1.38 |
Agentic & Tool Use Claude Haiku 4.5 leads
Claude Haiku 4.5: 33.5 (#50), GPT-5 Mini: 31.0 (#64)
| Benchmark | Claude Haiku 4.5 | GPT-5 Mini |
|---|---|---|
| Terminal-Bench | 35.5% | 34.8% |
| Berkeley Function Calling Leaderboard | 68.7% | 55.5% |
| Vending-Bench 2 | 458.89 | -31.18 |
| DeepResearch Bench | 45.5% | — |
| BALROG | 31.2% | — |
| ExploitBench | 13.7% | — |
Reasoning GPT-5 Mini leads
Claude Haiku 4.5: 15.1 (#310), GPT-5 Mini: 24.1 (#153)
| Benchmark | Claude Haiku 4.5 | GPT-5 Mini |
|---|---|---|
| ARC-AGI-2 | 4% | 4.4% |
| ARC-AGI-1 | 47.7% | 54.3% |
| CritPt | 0% | 0% |
| Chess Puzzles | 8% | 30% |
| LMArena Hard Prompts | 1420 | 1380 |
| DTBench | 73.6% | 80.5% |
| LMCA | 30.9% | 34.2% |
| Epoch Capabilities Index | 142.42 | 145.51 |
| ForecastBench | 61.4 | 61 |
| Kagi LLM Benchmark | — | 70.3% |
| NYT Connections (extended) | 14.3% | — |
| EnigmaEval | — | 8.2% |
| Mystery Game Puzzles | — | 10% |
Math GPT-5 Mini leads
Claude Haiku 4.5: 43.0 (#69), GPT-5 Mini: 46.3 (#59)
| Benchmark | Claude Haiku 4.5 | GPT-5 Mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.7% | 86.7% |
| Omni-MATH | 56.1% | 72.2% |
| LMArena Math | 1395 | 1377 |
| MATH Level 5 | 96.4% | 97.8% |
| FrontierMath (Feb 2025 set) | 5.9% | 27.2% |
| FrontierMath Tier 4 (v1) | 2.1% | 6.3% |
| FrontierMath (Tiers 1-3) | — | 46.7% |
| FrontierMath Tier 4 | — | 12.2% |
| ProofBench | — | 9% |
Knowledge GPT-5 Mini leads
Claude Haiku 4.5: 37.6 (#146), GPT-5 Mini: 45.5 (#80)
| Benchmark | Claude Haiku 4.5 | GPT-5 Mini |
|---|---|---|
| GPQA Diamond | 71.2% | 75% |
| SimpleQA Verified | 13.2% | 21.6% |
| MMLU-Pro | 77.7% | 83.5% |
| Vectara Hallucination Rate | 9.8% | 12.9% |
| GPQA (HELM) | 60.5% | 75.6% |
| LMArena Expert | 1443 | 1377 |
| Humanity's Last Exam | — | 19.4% |
| Confabulations | — | 13.3% |
Multimodal GPT-5 Mini leads
Claude Haiku 4.5: 27.0 (#111), GPT-5 Mini: 35.0 (#80)
| Benchmark | Claude Haiku 4.5 | GPT-5 Mini |
|---|---|---|
| LMArena Vision | — | 1202 |
| VPCT | — | 40.2% |
| Blueprint-Bench 2 | 0% | — |
| LMArena Document | 1420 | — |
Multilingual Too close to call
Claude Haiku 4.5: 50.1 (#124), GPT-5 Mini: 49.1 (#131)
| Benchmark | Claude Haiku 4.5 | GPT-5 Mini |
|---|---|---|
| LMArena Non-English | 1376 | 1363 |
| LMArena Chinese | 1415 | 1387 |
| LMArena French | 1408 | 1385 |
| LMArena German | 1377 | 1366 |
| LMArena Japanese | 1339 | 1342 |
| LMArena Korean | 1346 | 1306 |
| LMArena Russian | 1380 | 1362 |
| LMArena Spanish | 1420 | 1355 |
Instruction Following GPT-5 Mini leads
Claude Haiku 4.5: 71.6 (#144), GPT-5 Mini: 76.4 (#43)
| Benchmark | Claude Haiku 4.5 | GPT-5 Mini |
|---|---|---|
| IFEval | 80.1% | 92.7% |
| LMArena Instruction Following | 1414 | 1357 |
Long Context Claude Haiku 4.5 leads
Claude Haiku 4.5: 43.6 (#86), GPT-5 Mini: 41.9 (#127)
| Benchmark | Claude Haiku 4.5 | GPT-5 Mini |
|---|---|---|
| LMArena Longer Query | 1427 | 1355 |
| Fiction.LiveBench | — | 69.4% |
Writing & Preference Claude Haiku 4.5 leads
Claude Haiku 4.5: 58.1 (#119), GPT-5 Mini: 55.4 (#143)
| Benchmark | Claude Haiku 4.5 | GPT-5 Mini |
|---|---|---|
| LMArena Text | 1396 | 1373 |
| LMArena Creative Writing | 1371 | 1325 |
| WildBench | 83.9% | 85.5% |
| LMArena Multi-Turn | 1409 | 1363 |
| Short-Story Creative Writing | — | 83.1% |
| EQ-Bench Creative Writing | — | 1313 |
| EQ-Bench 4 | 1064 | — |
Frequently asked questions
Is Claude Haiku 4.5 better than GPT-5 Mini?
GPT-5 Mini is the stronger model overall, scoring 41.7 to 39.3 on the Noometry Index.
Which is cheaper, Claude Haiku 4.5 or GPT-5 Mini?
GPT-5 Mini is cheaper. It lists at $0.25 per million input tokens and $2 per million output tokens; Claude Haiku 4.5 lists at $1 and $5.
Is Claude Haiku 4.5 or GPT-5 Mini better for coding?
Claude Haiku 4.5 scores higher on coding benchmarks: 43.9 versus 40.0 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Mini does, with 400K tokens against 200K.
How many benchmarks do Claude Haiku 4.5 and GPT-5 Mini share?
45 benchmarks have published results for both models. Claude Haiku 4.5 has 53 scored results on Noometry and GPT-5 Mini has 60.