Kimi K3vsClaude Opus 4.8
Across 6 shared benchmarks, Kimi K3 leads overall: Kimi K3 wins 4, Claude Opus 4.8 wins 2, with 0 ties and an average score difference of +4.13.
Kimi K3
Moonshot AI · 2026-07-16 · Reasoning model
Claude Opus 4.8
Anthropic · 2026-05-28 · Reasoning model
Kimi K34 wins(67%)(33%)2 winsClaude Opus 4.8
Benchmark scores
Grouped by capability, sorted by largest gap within each. 6 shared benchmarks.
AI Agent - Tool Usage
Kimi K3 2/2| Benchmark | Kimi K3 | Claude Opus 4.8 | Diff |
|---|---|---|---|
| TerminalBench 2.1 | 88.302 / 25Max (With Tools) | 78.9012 / 25Thinking High (With Tools) | +9.40 |
| MCP-Atlas | 84.202 / 27Max (With Tools) | 82.205 / 27Deep Thinking (With Tools) | +2 |
General Knowledge
Claude Opus 4.8 2/2| Benchmark | Kimi K3 | Claude Opus 4.8 | Diff |
|---|---|---|---|
| HLE | 5610 / 170Max (With Tools) | 57.906 / 170Extended (with tools) | -1.90 |
| GPQA Diamond | 93.508 / 187最高（无工具） | 93.606 / 187Thinking High (No Tools) | -0.10 |
AI Agent - Information Search
Kimi K3 1/1| Benchmark | Kimi K3 | Claude Opus 4.8 | Diff |
|---|---|---|---|
| BrowseComp | 91.201 / 52Max (With Tools + Internet) | 84.308 / 52Thinking High (With Tools + Internet) | +6.90 |
Coding and Software Engineer
Kimi K3 1/1| Benchmark | Kimi K3 | Claude Opus 4.8 | Diff |
|---|---|---|---|
| DeepSWE | 67.504 / 17Max (With Tools) | 597 / 17Deep Thinking (With Tools) | +8.50 |
Specs
| Field | Kimi K3 | Claude Opus 4.8 |
|---|---|---|
| Publisher | Moonshot AI | Anthropic |
| Release date | 2026-07-16 | 2026-05-28 |
| Model type | Reasoning model | Reasoning model |
| Architecture | MoE | Dense |
| Parameters | 2.8T | Not available |
| Context length | 1M | 1M |
| Max output | 1M | 125K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | Kimi K3 | Claude Opus 4.8 |
|---|---|---|
| Text input | ¥20 / 1M tokens | $5 / 1M tokens |
| Text output | ¥100 / 1M tokens | $25 / 1M tokens |
| Cache read | ¥2 / 1M tokens | $0.5 / 1M tokens |
| Cache write | Not public | $6.25 / 1M tokens |
Summary
- Kimi K3leads in:AI Agent - Tool Usage (2/2), AI Agent - Information Search (1/1), Coding and Software Engineer (1/1)
- Claude Opus 4.8leads in:General Knowledge (2/2)
On average across the 6 shared benchmarks, Kimi K3 scores 4.13 higher.
Largest single-benchmark gap: TerminalBench 2.1 — Kimi K3 88.30 vs Claude Opus 4.8 78.90 (+9.40).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.