Across 18 comparable benchmarks, Claude Opus 4 has better reported results on 16 and worse on 2.
Moderate evidence18 of 18 shared results countedHow we compare
Claude Opus 4 is no longer available through its official API. API prices for this version are not shown.
Price basis: Standard rates vs listed rates. The price section below shows the applicable periods.
| Fact | GPT-4.1 | Claude Opus 4 |
|---|---|---|
| Context window | 1M tokens | 200K tokens |
| Input / 1M | $2.00 | — |
| Output / 1M | $8.00 | — |
| Blended (4:1) | $2.12 | — |
| Status | Deprecated | Retired |
| Released | April 14, 2025 | May 23, 2025 |
Blended is 4:1 input to output with 90% cache hits at listed rates — set your own mix.
Mean of 12 shared benchmarks grouped by capability on a 0–100 scale. How averages are built
Claude Opus 4 ahead on 4 of 4 counted
Claude Opus 4 ahead on 1 of 1 counted
No fixed-scale average — see the rows in Benchmark results
Claude Opus 4 ahead on 4 of 5 counted
Claude Opus 4 ahead on 1 of 1 counted
Claude Opus 4 ahead on 1 of 1 counted
GPT-4.1 ahead on 1 of 1 counted
Claude Opus 4 ahead on 2 of 2 counted
No shared comparable results for Computer & tool use, Multimodal, Cybersecurity.
Published scores, with matching modes where available. Gaps show GPT-4.1 relative to Claude Opus 4. Reading the results
| Benchmark | GPT-4.1 | Claude Opus 4 | Gap |
|---|---|---|---|
| Aider-PolyglotAgentic Development · Accuracy | 52.4Normal · no tools | 72.0≠Thinking · no tools · 32K | −19.6 pts#22 vs #9 of 47 |
| LiveCodeBenchAlgorithmic Coding · Pass @K | 40.5Normal · no tools | 56.6Normal · no tools | 0.72×#93 vs #75 of 113 |
| SWE-bench VerifiedRepository Engineering · Accuracy | 54.6Normal · no tools | 72.5Normal · no tools | −17.9 pts#85 vs #52 of 108 |
| Terminal Bench HardAgentic Development · Accuracy | 13.6Normal · tools | 31.1≠Thinking · tools | −17.5 pts#110 vs #66 of 157 |
| SimpleBenchCommonsense · Score (AVG@5) | 27.0Normal · no tools | 58.8≠Thinking · no tools | 0.46×#91 vs #51 of 106 |
| AIME 2024Mathematics · Accuracy | 48.1Normal · no tools | 76.0Normal · no tools | −27.9 pts#44 vs #31 of 54 |
| AIME2025Mathematics · Accuracy | 36.7Normal · no tools | 75.5Normal · no tools | −38.8 pts#82 vs #62 of 86 |
| FrontierMathMathematics · Accuracy | 5.5Normal · no tools | 4.5Normal · no tools | +1.0 pts#28 vs #30 of 48 |
| FrontierMath - Tier 4Mathematics · Accuracy | 0.0Normal · no tools | 4.2≠Thinking · no tools · 32K | −4.2 pts#44 vs #21 of 50 |
| MATH-500Mathematics · Accuracy | 92.8Normal · no tools | 98.2≠Thinking · no tools | −5.4 pts#38 vs #12 of 49 |
| GPQA DiamondScientific Reasoning · Accuracy | 66.3Normal · no tools | 70.1Normal · no tools | −3.8 pts#152 vs #140 of 200 |
| GAIAMulti-step Synthesis · Accuracy | 49.7Normal · tools | 64.8≠Thinking High · tools | −15.1 pts#8 vs #3 of 11 |
| Fiction.liveBenchLong Reasoning · 16k-token score (%) | 63.9Normal · no tools | 61.1Normal · no tools | +2.8 pts#17 vs #19 of 28 |
| Creative WritingWriting | 1420Normal · no tools | 1578Normal · no tools | 0.90×#71 vs #48 of 121 |
| τ²-BenchService Workflows · Accuracy | 54.7Normal · tools | 72.5≠Thinking · tools | −17.8 pts#30 vs #24 of 40 |
| HLEKnowledge Exams · Accuracy | 5.4Reported best | 10.7≠Thinking · no tools | −5.3 pts#141 vs #125 of 151 |
| MMLU-ProKnowledge Exams · Accuracy | 80.5Normal · no tools | 87.3≠Thinking · no tools | −6.8 pts#59 vs #13 of 153 |
| ECICapability Indices · ECI score (capability index, higher is better) | 137Reported best | 143Reported best | 0.96×#108 vs #87 of 168 |
Agentic Development · Accuracy
Algorithmic Coding · Pass @K
Repository Engineering · Accuracy
Agentic Development · Accuracy
Commonsense · Score (AVG@5)
Mathematics · Accuracy
Mathematics · Accuracy
Mathematics · Accuracy
Mathematics · Accuracy
Mathematics · Accuracy
Scientific Reasoning · Accuracy
Multi-step Synthesis · Accuracy
Long Reasoning · 16k-token score (%)
Writing
Service Workflows · Accuracy
A dash means no result is available.
| Benchmark | GPT-4.1 | Claude Opus 4 |
|---|---|---|
| FrontierMath v2 | 6.0 | — |
| GSM8K | 95.9 | — |
| IMO-ProofBench | — | 2.9 |
| IMO-ProofBench Advanced | — | 2.9 |
| MMLU | 90.2 | — |
| IC SWE-Lancer (Diamond) | 14.4 | — |
| SWE Manager Lancer (Diamond) | 35.1 | — |
| WeirdML v2 | 39.0 | — |
| DROP | 89.2 | — |
| GeoBench ACW | 72.0 | — |
| MMMU | — | 76.5 |
| MMMU-Pro | 61.2 | — |
| VPCT | — | 38.0 |
| τ²-Bench - Telecom | 34.0 | — |
| ARC-AGI-1 | — | 35.7 |
| ARC-AGI-2 | — | 8.6 |
Time-of-day providers use peak-hour rates; fixed-price providers use standard rates. Off-peak discounts remain in the published details.
Drag the handle above the bar for cache hits, the one below it for the input : output mix · 5% steps
Blended price = effective input rate at 90% cache hits × 80% + output rate × 20%
A blended cost needs a published input and output price on both sides, and one of these two does not have them yet.
Claude Opus 4: Listed text rates; pricing tier not confirmed.
| Price / 1M tokens | GPT-4.1 | Claude Opus 4 | Ratio |
|---|---|---|---|
| Input | $2.00 | — | — |
| Output | $8.00 | — | — |
| Cached input | $0.50 | — | — |
| Cache write | — | — | — |
| Blended | $2.12 | — | — |
Estimates use the selected token mix. Cache writes, tool fees and long-context surcharges are excluded.
| Attribute | GPT-4.1 | Claude Opus 4 |
|---|---|---|
| Availability | Deprecated | Retired |
| Provider | OpenAI | Anthropic |
| Model type | Chat model | Reasoning model |
| First release | April 14, 2025 | May 23, 2025 |
| Context window | 1M tokens | 200K tokens |
| Max output | 32K tokens | 32K tokens |
| Architecture | Undisclosed | Undisclosed |
| Documented capabilities | Multilingual | Reasoning, Multilingual |
| Input modalities | Text, Image | Text, Image |
| Output modalities | Text | Text |
| Weights | Closed / API only | Closed / API only |
| Commercial use | Proprietary | Proprietary |
These are reported-score comparisons, not controlled head-to-head tests.
What people found comparing these models, or using either one. Comments also appear on the models' own pages.
Used GPT-4.1 or Claude Opus 4? Sign in with Google to share what worked, what did not, and which API or setup you tested.
No community comments on either model yet.
Alternatives with the most benchmark overlap for this matchup.