Kimi K2
Balancedby Moonshot (Kimi)
Strong long-context reasoning.
At a glance
- Provider
- Moonshot (Kimi)
- Best for
- Balanced
- Reads images
- Yes
- Price
- from $10 / 1M tokens
- Subscription
- None — pay as you go
Benchmarks
Higher is better.
- GPQA Diamond· graduate-level reasoning
- 90.5%
- SWE-Bench Verified· real-world coding
- 80.2%
Source: llm-stats.com, as of 28 Jun 2026. Figures are publisher-reported where not independently verified.
Use Kimi K2 without a subscription
On the GPQA Diamond graduate-level reasoning benchmark, Kimi K2 scores 90.5%. anyAInow gives you Kimi K2 alongside every other top AI model from a single balance — from $10 / 1m tokens, with no monthly fee and tokens that never expire. Start a chat on Kimi K2, switch to another model mid-conversation, or run two models side by side to compare their answers. Top up from $5 and only pay for what you use.
Other models
GPT-5.4
A balanced mix of power and cost.
Claude Sonnet 4.6
The most popular Claude model — a strong all-rounder.
Gemini 3.5 Flash
Fast multimodal tasks.
Grok 4.20
Proven, balanced reasoning.
Mistral Medium
Strong coding and analysis.
Llama 4 Maverick
Meta's most capable open model.