Claude Opus 4.5
Claude Opus 4.5 (Anthropic) was released on Nov 24, 2025. In the Epoch AI capability ranking (ECI), it ranks #55 out of 267 models, with 150.17. On Arena, which measures human preference in blind comparisons, it has 1,504 ELO points (#18). On LiveBench, the overall average is 72.6, #41 out of 58 scored models. Against models of the same generation, its best result is on SWE-bench Verified: 76.7%, #2 of 19.
Where Claude Opus 4.5 stands out and where it falls behind, by the numbers
Claude Opus 4.5 has published results on 39 tests compiled by Epoch AI. The yardstick here is its generation: its position among the models released between Feb 27, 2025 and Feb 22, 2026 that took the same test. Against the whole archive, which includes 2023 models, a recent release lands near the top of almost everything, and that says nothing.
Where it stands out
Tests where it sits in the top third of its own generation.
Where it falls behind
Tests where it sits in the bottom half of its own generation.
See all 39 tests, one by one
| Test | Score | In its generation | Overall | Test record |
|---|---|---|---|---|
| GPQA Diamond | 86.0% | #13 / 56 | #63 / 173 | 96.4% Gemini 4 Argon |
| AIME 2024 and 2025 (OTIS) | 86.1% | #16 / 57 | #73 / 168 | 100.0% GPT-5.5 |
| Dtbench | 89.9% | #7 / 51 (tied with 1) | #42 / 148 | 98.9% Claude Opus 5.5 |
| Lmca | 44.46 | #4 / 46 | #36 / 121 | 68.23 Claude Opus 5.5 |
| WeirdML | 63.7% | #7 / 46 | #24 / 112 | 93.6% GPT-6 Astra |
| Frontiermath Tiers 1 3 V2 | 34.4% | #11 / 20 | #62 / 86 | 93.7% GPT-6 Astra |
| ARC-AGI 2 | 37.6% | #7 / 40 | #38 / 83 | 95.0% GPT-6 Astra |
| ALE-Bench | 1,025 | #9 / 33 | #38 / 82 | 2,951 GPT-6 Astra |
| WebDev Arena | 1,495 | #3 / 24 | #43 / 82 | 1,820 Claude Opus 5.5 |
| ARC-AGI | 80.0% | #8 / 40 | #39 / 81 | 98.5% Claude Fable 5 |
| SimpleQA Verified | 45.7% | #8 / 23 | #38 / 81 | 76.0% Gemini 4 Argon |
| Chess puzzles | 14.0% | #16 / 22 | #65 / 80 | 76.0% GPT-6 Astra |
| Game puzzles | 22.0% | #7 / 20 | #39 / 77 | 84.0% GPT-6 Astra |
| ProofBench | 36.0% | #3 / 15 | #38 / 74 | 100.0% Claude Fable 5.1 |
| SimpleBench | 62.0% | #5 / 32 | #18 / 74 | 81.9% Claude Fable 5 |
| Frontiermath Tier 4 V2 | 4.9% | #10 / 15 (tied with 1) | #57 / 67 | 100.0% GPT-6.1 Sol |
| FrontierMath | 20.7% | #15 / 37 | #29 / 65 | 51.7% GPT-5.5 |
| ForecastBench | 60.70 | #16 / 32 | #20 / 62 | 62.50 o3 |
| Vending Bench 2 | 4,967 | #5 / 23 | #32 / 57 | 15,515 GPT-6 Astra |
| FrontierMath tier 4 | 4.2% | #12 / 31 (tied with 6) | #23 / 48 | 39.6% GPT-5.5 Pro |
| Hle | 25.2% | #7 / 22 | #16 / 39 | 54.8% GPT-6 Astra |
| EnigmaEval | 11.9% | #5 / 21 | #11 / 36 | 39.3% Claude Fable 5 |
| Furniture Assembly | 28.3% | #2 / 4 (tied with 1) | #25 / 34 | 83.3% Claude Opus 5.5 |
| Terminal-Bench | 63.1% | #7 / 27 | #10 / 33 | 84.7% GPT-5.5 |
| BALROG | 43.5% | #5 / 11 | #10 / 32 | 68.3% GPT-6 Astra |
| SWE-bench Verified | 76.7% | #2 / 19 | #9 / 32 | 83.5% Claude Opus 4.7 |
| METR: time horizon | 293 | #5 / 19 | #6 / 29 | 719 Claude Opus 4.6 |
| Gso | 26.5% | #3 / 15 | #12 / 28 | 88.2% Claude Fable 5.1 |
| Ebr Bench | 14.3% | #2 / 7 (tied with 1) | #18 / 27 | 76.2% GPT-6 Astra |
| DeepResearch Bench | 54.8% | #3 / 18 | #3 / 24 | 55.3% Claude Opus 4.6 |
| GeoBench | 3,877 | #6 / 14 | #9 / 24 | 4,333 Gemini 3 Flash |
| VPCT | 40.0% | #12 / 17 | #12 / 22 | 91.0% Gemini 3 Pro |
| Cl Bench | 21.1% | #3 / 17 | #4 / 19 | 27.9% GPT-5.4 |
| Cybench | 82.0% | #2 / 9 | #2 / 19 | 93.0% Claude Opus 4.6 |
| Algotune | 1.77 | #4 / 14 | #5 / 15 | 2.05 GPT-5.2 |
| Rli | 3.8% | #2 / 7 | #10 / 15 | 21.3% Claude Opus 5.5 |
| GDPval | 45.5% | #2 / 10 | #2 / 11 | 49.7% GPT-5.2 |
| Weirdml V3 | 6.5% | #1 / 1 | #10 / 11 | 42.2% GPT-6 Astra |
| Os World | 66.30 | #2 / 6 | #2 / 8 | 72.10 Claude Sonnet 4.6 |
For each test, the model's best effort level counts, as in the ranking. Source of the scores: Epoch AI, under CC BY 4.0.
What to compare Claude Opus 4.5 with
The models from other labs closest in capability and released in the same period, at most two per lab. Click any of them to see its page.
| Model | Lab | ECI | ELO | LiveBench | Release |
|---|---|---|---|---|---|
| Kimi K2.6 | Moonshot | 150.92 | 1,521 | 70.5 | Apr 20, 2026 |
| Inkling-Small | Thinking Machines | 150.18 | 1,463 | n/a | Jul 15, 2026 |
| Claude Opus 4.5 | Anthropic | 150.17 | 1,504 | 72.6 | Nov 24, 2025 |
| GPT-5 | OpenAI | 150.00 | 1,436 | n/a | Aug 7, 2025 |
| GLM-5.1 | Z.ai (Zhipu AI) | 149.99 | 1,514 | n/a | Apr 7, 2026 |
| Kimi K2.7 Code | Moonshot | 149.87 | n/a | 68.4 | Jun 12, 2026 |
| GPT-5.1 | OpenAI | 149.76 | 1,495 | n/a | Nov 13, 2025 |
Claude Opus 4.5 or Inkling-Small?
Inkling-Small is the model from another lab closest to Claude Opus 4.5 in capability. Across the 10 tests both took, this is the score:
| Test | Claude Opus 4.5 | Inkling-Small |
|---|---|---|
| ProofBench | 36.0% | 6.0% |
| Game puzzles | 22.0% | 6.0% |
| Frontiermath Tier 4 V2 | 4.9% | 17.1% |
| SimpleQA Verified | 45.7% | 19.1% |
| Frontiermath Tiers 1 3 V2 | 34.4% | 46.3% |
| ARC-AGI 2 | 37.6% | 40.1% |
| WebDev Arena | 1,495 | 1,407 |
| ARC-AGI | 80.0% | 84.0% |
| AIME 2024 and 2025 (OTIS) | 86.1% | 90.0% |
| GPQA Diamond | 86.0% | 88.5% |
Unweighted score: each test counts one point, whether math or coding. The dot marks who did better on each row.
What Claude Opus 4.5 costs to deliver each task
Price per token is what is on the price sheet; cost per task is what lands on the invoice, because it depends on how many tokens the model spends. Each point is an effort level, published by whoever ran the test and compiled by Epoch AI. The further up and to the left, the more score per dollar.
ARC-AGI 2
The score scale does not start at zero.
ARC-AGI
The score scale does not start at zero.
Other models from Anthropic
The models from the same lab closest in time, newest first.
| Model | Release | ECI | ELO | LiveBench | Access |
|---|---|---|---|---|---|
| Claude Opus 4.8 | May 28, 2026 | 158.26 | 1,507 | 76.2 | API access |
| Claude Opus 4.7 | Apr 16, 2026 | 156.32 | 1,532 | 76.5 | API access |
| Claude Sonnet 4.6 | Feb 17, 2026 | 152.39 | 1,504 | 73.0 | API access |
| Claude Opus 4.6 | Feb 5, 2026 | 155.27 | 1,550 | 74.5 | API access |
| Claude Opus 4.5 | Nov 24, 2025 | 150.17 | 1,504 | 72.6 | API access |
| Claude Haiku 4.5 | Oct 15, 2025 | 142.40 | 1,454 | n/a | API access |
| Claude Sonnet 4.5 | Sep 29, 2025 | 146.83 | 1,488 | n/a | API access |
| Claude Opus 4.1 | Aug 5, 2025 | 144.09 | 1,479 | n/a | API access |
| Claude Opus 4 | May 22, 2025 | 142.65 | 1,442 | n/a | API access |
Have you used Claude Opus 4.5? Rate it
Users' opinion, not measurement: the rating is kept apart from the ranking and does not move anyone's position. One vote per browser, no sign-up, and you can change your rating later.
What people ask about Claude Opus 4.5
Where does Claude Opus 4.5 rank among AI models?
In the ROO3 AI Benchmark, ordered by the Epoch AI ECI, Claude Opus 4.5 ranks #55 out of 267 models, with an ECI of 150.17. The ECI is a composite index that combines several hard benchmarks into a single scale. The position changes when a new model comes out, and the page is refreshed six times a day.
Is Claude Opus 4.5 better than Inkling-Small?
Across the 10 tests both models took and Epoch AI compiled, Claude Opus 4.5 comes out ahead on 4, behind on 6 and tied on 0. It is a test-by-test count with no weighting: a win in math counts the same as a win in coding. To choose between the two, look at the tests that resemble your own use.
What is Claude Opus 4.5 best at?
Against models of the same generation (released up to 9 months before and up to 3 months after it), the best relative results of Claude Opus 4.5 are: SWE-bench Verified (#2 of 19); Lmca (#4 of 46); WebDev Arena (#3 of 24). The yardstick is its position among the models of that period that took the same test.
Where do the numbers on this page come from?
From the same three sources as the ranking, all openly licensed: Epoch AI (ECI and individual tests, CC BY 4.0), Arena (human-preference ELO, CC BY 4.0) and LiveBench (overall average, CC BY-SA 4.0). ROO3 does not measure any model. When the page shows a figure published by the maker itself, it appears in the written analysis, labeled as the maker's, and never enters the ranking.
Where each number comes from
Epoch AI, "AI Benchmarking Hub". Published at epoch.ai. Data under the CC BY 4.0 licence, reordered by roo3.co.
The Arena (LMArena) leaderboard, from the public lmarena-ai/leaderboard-dataset on Hugging Face, under the CC BY 4.0 licence. We are not affiliated with the Arena.
LiveBench, by the LiveBench team, published at livebench.ai under the CC BY-SA 4.0 license. The overall average is recomputed by roo3.co with the same formula as the website (average of the categories), and those numbers stay under CC BY-SA 4.0. We are not affiliated with LiveBench.
← See the full ranking