Compare / head-to-head
Claude Opus 4.8vs
Gemini 2.5 Pro
Claude Opus 4.8 leads 14 of 15 shared benchmarks. Gemini 2.5 Pro is 2.7x cheaper per token. Gemini 2.5 Pro has the larger context window (1.0M tokens).
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
14 – 1
Claude Opus 4.8 leads
Cheaper per token
Gemini 2.5 Pro
2.7x cheaper, input + output
Larger context
Gemini 2.5 Pro
1.0M tokens
Providers
2 providers
Anthropic · Google
Claude Opus 4.8
Anthropic · released 2026-05-28
textvision
Gemini 2.5 Pro
Google · released 2025-06-17
textvisionaudio
Quality
Benchmark matrix
15 shared · 25 only Claude Opus 4.8 · 5 only Gemini 2.5 Pro| Benchmark | Claude Opus 4.8 | Gemini 2.5 Pro | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 15 | ||||
| SWE-bench Verified | 88.6% ↗ | 59.6% ↗ | +29 pt | Claude Opus 4.8 |
| GPQA Diamond | 93.6% ↗ | 86.4% ↗ | +7.2 pt | Claude Opus 4.8 |
| Humanity's Last Exam (no tools) | 49.8% ↗ | 21.6% ↗ | +28.2 pt | Claude Opus 4.8 |
| LMArena Elo | 1452.7 ↗ | 1457.8 ↗ | -5.1 | Gemini 2.5 Pro |
| Chess Puzzles (Epoch AI run) | 34% ↗epoch run | 20% ↗epoch run | +14 pt | Claude Opus 4.8 |
| FrontierMath Tier 4 v2 (Epoch AI run) | 56.1% ↗epoch run | 0% ↗epoch run | +56.1 pt | Claude Opus 4.8 |
| FrontierMath Tiers 1-3 v2 (Epoch AI run) | 80% ↗epoch run | 24.6% ↗epoch run | +55.4 pt | Claude Opus 4.8 |
| GSO Opt@1 (GSO) | 47.1% ↗ | 0% ↗ | +47.1 pt | Claude Opus 4.8 |
| LMArena Vision (LMArena) | 1286.4 ↗ | 1247.5 ↗ | +38.9 | Claude Opus 4.8 |
| LMArena WebDev (LMArena) | 1555.5 ↗ | 1227 ↗ | +328.5 | Claude Opus 4.8 |
| OTIS Mock AIME 2024-2025 (Epoch AI run) | 98.3% ↗epoch run | 84.7% ↗epoch run | +13.6 pt | Claude Opus 4.8 |
| SAGE (Vals AI) | 54.8% ↗ | 41.9% ↗ | +12.9 pt | Claude Opus 4.8 |
| tau2-bench Banking Knowledge (Sierra) | 39.7% ↗ | 13.7% ↗ | +26 pt | Claude Opus 4.8 |
| Vending-Bench 2 (Andon Labs) | 5787.43 ↗ | 573.64 ↗ | +5213.8 | Claude Opus 4.8 |
| WeirdML (Håvard Tveit Ihle) | 82.9% ↗ | 54% ↗ | +28.9 pt | Claude Opus 4.8 |
| Only Claude Opus 4.8 reports · 25 | ||||
| AIME 2026 | 100% ↗matharena ⚠ | not reported | — | — |
| APEX-Agents (Mercor) | 48.9% ↗ | not reported | — | — |
| BrowseComp | 84.3% ↗ | not reported | — | — |
| DeepSWE v1.1 (Datacurve) | 59% ↗ | not reported | — | — |
| EBR-bench (Epoch AI run) | 28.6% ↗epoch run | not reported | — | — |
| Furniture Assembly (Epoch AI run) | 42.5% ↗epoch run | not reported | — | — |
| Humanity's Last Exam (with tools) | 57.9% ↗ | not reported | — | — |
| LiveBench Agentic Coding (LiveBench) | 50.5% ↗ | not reported | — | — |
| LiveBench Coding (LiveBench) | 81.8% ↗ | not reported | — | — |
| LiveBench Data Analysis (LiveBench) | 66% ↗ | not reported | — | — |
| LiveBench Instruction Following (LiveBench) | 72% ↗ | not reported | — | — |
| LiveBench Language (LiveBench) | 79.7% ↗ | not reported | — | — |
| LiveBench Mathematics (LiveBench) | 94.3% ↗ | not reported | — | — |
| LiveBench Reasoning (LiveBench) | 89.2% ↗ | not reported | — | — |
| LMArena Agent (LMArena) | 0.0664 ↗ | not reported | — | — |
| Mystery Game Puzzles (Epoch AI run) | 36% ↗epoch run | not reported | — | — |
| OSWorld-Verified | 83.4% ↗ | not reported | — | — |
| SimpleBench (SimpleBench) | 64.8% ↗ | not reported | — | — |
| SimpleQA Verified | 53% ↗epoch run | not reported | — | — |
| SWE-bench Multilingual | 84.4% ↗ | not reported | — | — |
| SWE-bench Multimodal | 38.4% ↗ | not reported | — | — |
| SWE-Bench Pro | 69.2% ↗ | not reported | — | — |
| Terminal-Bench 2.1 | 74.6% ↗ | not reported | — | — |
| Terminal-Bench 4.0 (Vals AI) | 23.2% ↗ | not reported | — | — |
| Toolathlon-Verified (HKUST) | 76.2% ↗ | not reported | — | — |
| Only Gemini 2.5 Pro reports · 5 | ||||
| AIME 2025 | not reported | 88% ↗ | — | — |
| Aider Polyglot | not reported | 82.2% ↗ | — | — |
| LiveCodeBench | not reported | 74.2% ↗ | — | — |
| MMMU | not reported | 82% ↗ | — | — |
| SWE-bench Verified (Epoch AI run) | not reported | 57.6% ↗epoch run | — | — |
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is Claude Opus 4.8 minus Gemini 2.5 Pro in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing
Side by side
Official pricing pages and model cards| Spec | Claude Opus 4.8 | Gemini 2.5 Pro | Edge |
|---|---|---|---|
| Context window | 1M tokens | 1.0M tokens | Gemini 2.5 Pro |
| Max output | 128K tokens | 66K tokens | Claude Opus 4.8 |
| Input price / 1M | $5 ↗ | $1.25 ↗ | Gemini 2.5 Pro |
| Output price / 1M | $25 ↗ | $10 ↗ | Gemini 2.5 Pro |
| Cached input / 1M | $0.5 ↗ | $0.125 ↗ | Gemini 2.5 Pro |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | $30 | $11.25 | Gemini 2.5 Pro |
| Modalities | text · vision | text · vision · audio | Gemini 2.5 Pro |
| Released | 2026-05-28 | 2025-06-17 | — |
| Cited benchmark scores | 47 | 28 | — |
Reliability
Provider status
Live from /statusMore matchups
Claude Opus 4.8 vs …
Models sharing the most benchmarksMore matchups
Gemini 2.5 Pro vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run Claude Opus 4.8 and Gemini 2.5 Pro on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.