Compare
Four models, one page
Pick the models you are actually choosing between. Every row is scaled within itself, so a bar means something next to its neighbours and nothing across rows.
| Specification | Gemini 2.5 ProGoogle | Claude Opus 5Anthropic |
|---|---|---|
| Released | Jun 17, 2025 | Jul 24, 2026 |
| Context window | 1.05M | 1M |
| Max output | 66K | 128K |
| Input / 1M | $1.25 | $5 |
| Output / 1M | $10 | $25 |
| Cost per task | $0.86 | $2.25 |
| Arena Elo | 1,273 | 1,393 |
| Serving providers | 2 | 5 |
| Parameters | Undisclosed | Undisclosed |
| Licence | Proprietary | Proprietary |
| Capabilities |
|
|
Metrics side by side
Each row is scaled to the largest value in that row — bars compare within a row, never across rows
- Gemini 2.5 Pro
- Claude Opus 5
Intelligence Index
Gemini 2.5 Pro25.9Claude Opus 563.1Coding Index
Gemini 2.5 Pro33.3Claude Opus 578.0Agentic Index
Gemini 2.5 Pro7.2Claude Opus 559.2Output speed
Gemini 2.5 Pro106 t/sClaude Opus 587 t/sContext window
Gemini 2.5 Pro1.05MClaude Opus 51MLatency · lower is better
Gemini 2.5 Pro2.32sClaude Opus 51.12sBlended price / 1M · lower is better
Gemini 2.5 Pro$3.44Claude Opus 5$10Cost per task · lower is better
Gemini 2.5 Pro$0.86Claude Opus 5$2.25
Rows marked “lower is better” still draw a longer bar for a larger number — read the value, not just the length. Arena Elo is in the specification table above instead: it has no meaningful zero, so a bar would flatten the gaps.
View as table
| Metric | Gemini 2.5 Pro | Claude Opus 5 |
|---|---|---|
| Intelligence Index | 25.9 | 63.1 |
| Coding Index | 33.3 | 78.0 |
| Agentic Index | 7.2 | 59.2 |
| Output speed | 106 t/s | 87 t/s |
| Context window | 1.05M | 1M |
| Latency · lower is better | 2.32s | 1.12s |
| Blended price / 1M · lower is better | $3.44 | $10 |
| Cost per task · lower is better | $0.86 | $2.25 |
Evaluation scores
Percentage correct on a common 0–100% scale
- Gemini 2.5 Pro
- Claude Opus 5
GPQA Diamond
Gemini 2.5 Pro84.4%Claude Opus 593.2%Humanity's Last Exam
Gemini 2.5 Pro22.5%Claude Opus 554.9%SciCode
Gemini 2.5 Pro42.8%Claude Opus 555.7%τ²-bench
Gemini 2.5 Pro54.1%Claude Opus 542.1%Terminal-Bench Hard
Gemini 2.5 Pro26.5%Claude Opus 5not measuredLiveCodeBench
Gemini 2.5 Pro80.1%Claude Opus 5not measuredAA-LCR long context
Gemini 2.5 Pro66.0%Claude Opus 575.7%AIME 2025
Gemini 2.5 Pro87.7%Claude Opus 5not measured
A missing bar means that evaluation was not run for that model — it is not a zero.
View as table
| Evaluation | Gemini 2.5 Pro | Claude Opus 5 |
|---|---|---|
| GPQA Diamond | 84.4% | 93.2% |
| Humanity's Last Exam | 22.5% | 54.9% |
| SciCode | 42.8% | 55.7% |
| τ²-bench | 54.1% | 42.1% |
| Terminal-Bench Hard | 26.5% | — |
| LiveCodeBench | 80.1% | — |
| AA-LCR long context | 66.0% | 75.7% |
| AIME 2025 | 87.7% | — |
Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).