AI Rankings
Claude Opus 4.8 vs GPT-5.5: Scores, Price and Which to Choose
Data updated
Read the model pages for the full picture: Claude Opus 4.8 and GPT-5.5.
Side by side
- Claude Opus 4.8
- GPT-5.5
- Epoch Capabilities IndexToo close to call
Combined score over many tests. Bar shows where it sits between the lowest and highest score we track.
Claude Opus 4.8:159.0GPT-5.5:159.7 - Arena text ratingToo close to call
Rating from head-to-head votes by people. Bar shows where it sits between the lowest and highest score we track.
Claude Opus 4.8:1482GPT-5.5:1482 - Arena coding ratingClaude Opus 4.8 leads
Rating from head-to-head votes by people. Bar shows where it sits between the lowest and highest score we track.
Claude Opus 4.8:Better: 1533GPT-5.5:1519 - Arena web development ratingClaude Opus 4.8 leads
Rating from head-to-head votes by people. Bar shows where it sits between the lowest and highest score we track.
Claude Opus 4.8:Better: 1556GPT-5.5:1513 - Arena vision ratingToo close to call
Rating from head-to-head votes by people. Bar shows where it sits between the lowest and highest score we track.
Claude Opus 4.8:1286GPT-5.5:1286 - DeepSWEGPT-5.5 leads
Software engineering tasks scored as the share solved on the first try (pass@1), run with the open mini-SWE-agent setup.
Claude Opus 4.8:59.0%GPT-5.5:Better: 67.0% - ARC-AGI-2GPT-5.5 leads
Visual puzzles where the AI must work out a new rule from a few examples and apply it.
Claude Opus 4.8:72.1%GPT-5.5:Better: 85.0% - SimpleBenchGPT-5.5 leads
Everyday common-sense and trick questions that most people answer correctly but AI models often get wrong.
Claude Opus 4.8:64.8%GPT-5.5:Better: 69.0% - FrontierMath (Tiers 1-3)GPT-5.5 leads
Original, very hard maths problems written by mathematicians.
Claude Opus 4.8:80.0%GPT-5.5:Better: 85.3% - GPQA DiamondGPT-5.5 leads
Very hard multiple-choice science questions in biology, physics and chemistry, written by experts and designed so a quick web search does not give the answer.
Claude Opus 4.8:91.0%GPT-5.5:Better: 94.0% - OTIS Mock AIMEToo close to call
Maths competition problems in the style of the US AIME exam, from the 2024 and 2025 sets.
Claude Opus 4.8:98.3%GPT-5.5:100.0% - SimpleQA VerifiedGPT-5.5 leads
Short factual questions that each have one correct answer.
Claude Opus 4.8:53.0%GPT-5.5:Better: 63.0% - Blended price per million tokensClaude Opus 4.8 costs less
Blended price: lower is better.
Claude Opus 4.8:Better: $10.00GPT-5.5:$11.25 - Context windowToo close to call
How much text it can read at once.
Claude Opus 4.8:1M tokensGPT-5.5:1.1M tokens
The numbers
| Measure | Claude Opus 4.8 | GPT-5.5 |
|---|---|---|
| Epoch Capabilities IndexToo close to call | 159.0 | 159.7 |
| Arena text ratingToo close to call | 1482 | 1482 |
| Arena coding rating | 1533 | 1519 |
| Arena web development rating | 1556 | 1513 |
| Arena vision ratingToo close to call | 1286 | 1286 |
| DeepSWE | 59.0% | 67.0% |
| ARC-AGI-2 | 72.1% | 85.0% |
| SimpleBench | 64.8% | 69.0% |
| FrontierMath (Tiers 1-3) | 80.0% | 85.3% |
| GPQA Diamond | 91.0% | 94.0% |
| OTIS Mock AIMEToo close to call | 98.3% | 100.0% |
| SimpleQA Verified | 53.0% | 63.0% |
| Blended price per million tokens | $10.00 | $11.25 |
| Context windowToo close to call | 1M tokens | 1.1M tokens |
What it costs in everyday terms
- 100 chats with Claude Opus 4.8
- $1.75
- Blended price $10.00 per million tokens.
- 100 chats with GPT-5.5
- $2
- Blended price $11.25 per million tokens.
Who should pick which
Pick Claude Opus 4.8 if these matter most
- Arena coding rating: 1533 against 1519
- Arena web development rating: 1556 against 1513
- Blended price per million tokens: $10.00 against $11.25
Pick GPT-5.5 if these matter most
- DeepSWE: 67.0% against 59.0%
- ARC-AGI-2: 85.0% against 72.1%
- SimpleBench: 69.0% against 64.8%
- FrontierMath (Tiers 1-3): 85.3% against 80.0%
- GPQA Diamond: 94.0% against 91.0%
- SimpleQA Verified: 63.0% against 53.0%
Too close to call: Epoch Capabilities Index, Arena text rating, Arena vision rating, OTIS Mock AIME, Context window.
Scores come from tests and votes, not from your own work. Try both on a task you care about before you commit.
Questions about this comparison
- Which is better, Claude Opus 4.8 or GPT-5.5?
- GPT-5.5 scores higher on 6 of 12 shared quality measures, Claude Opus 4.8 on 2. 4 measures are too close to call. Their blended prices are similar ($10.00 and $11.25 per million tokens). Which one is better for you depends on your task and budget, so test both on your own work.
- Which is cheaper, Claude Opus 4.8 or GPT-5.5?
- Claude Opus 4.8 is cheaper on blended price: $10.00 per million tokens for Claude Opus 4.8 against $11.25 for GPT-5.5. For 100 chats of about 1,000 tokens in and 500 tokens back, that is about $1.75 for Claude Opus 4.8 and $2 for GPT-5.5.
- Which is better for coding, Claude Opus 4.8 or GPT-5.5?
- On Arena coding rating, Claude Opus 4.8 scores higher (1533 against 1519). One measure is not the whole picture, so check the full table above.
Where this data comes from
- Epoch AI: Benchmark data: Epoch AI, 'Capabilities & benchmarking', epoch.ai/benchmarks (CC BY 4.0). We match models across sources, rank them and show them in charts. Scores are not changed. Epoch AI does not endorse this site.
- Arena (formerly LMArena): Human-preference ratings: Arena (arena.ai) leaderboard dataset, CC BY 4.0. We keep the latest published ratings, match them to models from other sources and rank them. Ratings are not changed. Arena does not endorse this site.
- LiteLLM model price list: Prices and context windows: LiteLLM model price list (github.com/BerriAI/litellm, MIT licence). We use the model maker's own price entry when there is one. Prices change often, so always confirm on the maker's site before you buy.
Newest data on this page: 26 Oct 2026. Prices are list prices and change often, so confirm on the maker's site before you buy. Sponsors never change a score or an order. How we calculate this · Report a mistake