Claude Opus 5vsKimi K3
Across 14 shared benchmarks, Claude Opus 5 leads overall: Claude Opus 5 wins 12, Kimi K3 wins 2, with 0 ties and an average score difference of +27.43.
Claude Opus 5
Anthropic · 2026-07-24 · Reasoning model
Kimi K3
Moonshot AI · 2026-07-16 · Reasoning model
Claude Opus 512 wins(86%)(14%)2 winsKimi K3
Benchmark scores
Grouped by capability, sorted by largest gap within each. 14 shared benchmarks.
Productivity Knowledge
Claude Opus 5 2/3| Benchmark | Claude Opus 5 | Kimi K3 | Diff |
|---|---|---|---|
| GDPval-AA v2 | 1,8611 / 26Max (With Tools) | 1,6869 / 26Max (With Tools) | +175 |
| AA-AnalystAgent | 53.752 / 12Max (With Tools + Internet) | 38.757 / 12Max (With Tools + Internet) | +15 |
| AutomationBench | 2612 / 15Max (With Tools) | 30.809 / 15Max (With Tools) | -4.80 |
AI Agent - Tool Usage
Claude Opus 5 2/2| Benchmark | Claude Opus 5 | Kimi K3 | Diff |
|---|---|---|---|
| Terminal-Bench-Science 0.1 | 303 / 11Max (With Tools) | 7.109 / 11Max (With Tools) | +22.90 |
| OSWorld 2.0 | 70.573 / 10Max (With Tools) | 58.308 / 10Max (With Tools) | +12.27 |
Coding and Software Engineer
Claude Opus 5 2/2| Benchmark | Claude Opus 5 | Kimi K3 | Diff |
|---|---|---|---|
| Text Arena (Coding) | 1,7121 / 35Max (No Tools) | 1,6822 / 35Max (No Tools) | +30.13 |
| DeepSWE | 68.808 / 35Max (With Tools) | 67.509 / 35Max (With Tools) | +1.30 |
Math and Reasoning
Claude Opus 5 2/2| Benchmark | Claude Opus 5 | Kimi K3 | Diff |
|---|---|---|---|
| FrontierMath Tier 4 v2 | 73.176 / 42Max (No Tools) | 39.0215 / 42Max (No Tools) | +34.15 |
| FrontierMath v2 | 85.615 / 58Max (No Tools) | 72.1813 / 58Max (No Tools) | +13.43 |
AI Agent - Information Search
Kimi K3 1/1| Benchmark | Claude Opus 5 | Kimi K3 | Diff |
|---|---|---|---|
| BrowseComp | 90.803 / 57Max (With Tools + Internet) | 91.202 / 57Max (With Tools + Internet) | -0.40 |
General Evaluation
Claude Opus 5 1/1| Benchmark | Claude Opus 5 | Kimi K3 | Diff |
|---|---|---|---|
| GPQA Diamond | 93.8813 / 273Max (No Tools) | 93.5017 / 273Max (No Tools) | +0.38 |
General Knowledge
Claude Opus 5 1/1| Benchmark | Claude Opus 5 | Kimi K3 | Diff |
|---|---|---|---|
| HLE | 64.702 / 191Max (With Tools) | 5617 / 191Max (With Tools) | +8.70 |
Text Embedding
Claude Opus 5 1/1| Benchmark | Claude Opus 5 | Kimi K3 | Diff |
|---|---|---|---|
| Context Arena | 97.721 / 126Max (No Tools) | 71.7559 / 126Max (No Tools) | +25.97 |
Writing and Creative Capabilities
Claude Opus 5 1/1| Benchmark | Claude Opus 5 | Kimi K3 | Diff |
|---|---|---|---|
| Creative Writing | 2,1213 / 106Normal (No Tools) | 2,0714 / 106Normal (No Tools) | +50 |
Specs
| Field | Claude Opus 5 | Kimi K3 |
|---|---|---|
| Publisher | Anthropic | Moonshot AI |
| Release date | 2026-07-24 | 2026-07-16 |
| Model type | Reasoning model | Reasoning model |
| Architecture | Dense | MoE |
| Parameters | Not available | 2.8T |
| Context length | 1M | 1M |
| Max output | 128K | 1M |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | Claude Opus 5 | Kimi K3 |
|---|---|---|
| Text input | $5 / 1M tokens | ¥20 / 1M tokens |
| Text output | $25 / 1M tokens | ¥100 / 1M tokens |
| Cache read | $0.5 / 1M tokens | ¥2 / 1M tokens |
| Cache write | $6.25 / 1M tokens | Not public |
Summary
- Claude Opus 5leads in:Productivity Knowledge (2/3), AI Agent - Tool Usage (2/2), Coding and Software Engineer (2/2), Math and Reasoning (2/2), General Evaluation (1/1), General Knowledge (1/1), Text Embedding (1/1), Writing and Creative Capabilities (1/1)
- Kimi K3leads in:AI Agent - Information Search (1/1)
On average across the 14 shared benchmarks, Claude Opus 5 scores 27.43 higher.
Largest single-benchmark gap: GDPval-AA v2 — Claude Opus 5 1,861 vs Kimi K3 1,686 (+175).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.