The Open-Source AI Stack
RSS
All models

Models · Compare

Claude Opus 4.6 vs Kimi K2.5

Rows highlighted in warm gray are where the models differ. Numbers carry their as-of date and primary source.

Specs

Field A: Claude Opus 4.6 B: Kimi K2.5
Released 2026-02-052026-01-27
Developer AnthropicMoonshot AI
Openness ProprietaryOpen
License ProprietaryModified MIT
OSI-approved nono
Data released nono
Training code nono
Architecture unknownmoe
Total params 1T
Active params 32B
Experts 384 (8 active)
Context window 1.0M256K
Attention unknownmla
Position enc. unknownrope
Pretraining tokens 15.0T
Post-training rlhf, constitutionalsft, rlhf
Training hardware
$/M input $5.00$0.58
$/M output $25.00$3.00
Output tok/sec 43.733.9

Benchmarks

Missing scores render as not reported; never inferred. Bold highlights the leader per benchmark.

General reasoning

GPQA-Diamond 84.0 2026-05-21 87.6 2026-01-27

Code

SWE-Bench Verified 76.8 2026-01-27
LiveCodeBench 85.0 2026-01-27

Math

AIME 2025 96.1 2026-01-27

Context · A

First Opus-class model with a 1M-token context window (in beta), released February 5 2026. Headline feature was agent teams in Claude Code, letting multiple agents work in parallel and coordinate autonomously. Long-context pricing premium kicks in above 200K input tokens ($10 / $37.50 per Mtok).

Context · B

Native multimodal Kimi at 1T total / 32B active with MoonViT (400M-parameter vision encoder), trained on roughly 15T mixed visual and text tokens. 256K context, Modified MIT license, January 27 2026 release. Adds visual reasoning, video understanding, and UI-to-code workflows.

Claude Opus 4.6 detail → · Kimi K2.5 detail →