Compare AI models

Specifications
Parameters
1T30B
Context window
256k200k
API pricing
Input price
—$0.00
Output price
—$0.00
Cached input price
—$0.00
Cheapest input
$0.60Novita—
Cheapest output
$2.50Novita—
$245Novita · bf16$0.00Z.ai list price
Benchmarks
SWE-Bench Verified
71.3%59.2%
LiveCodeBench
83.1%64%
BrowseComp
60.2%42.8%
AIME 2025
94.5%91.6%
Overview
CompanyMoonshot AIZ.ai
Release dateNov 6 2025Jan 19 2026
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Kimi K2 Thinking leads GLM-4.7-Flash on 4 of the 4 benchmarks they both report (SWE-Bench Verified, LiveCodeBench, BrowseComp, AIME 2025). Only GLM-4.7-Flash has a verified first-party price for an available API: $0.00 per million input tokens and $0.00 per million output tokens. No currently available pay-as-you-go API rate is tracked for Kimi K2 Thinking. Kimi K2 Thinking shipped 74 days before GLM-4.7-Flash, so benchmark comparisons should account for the intervening progress.

Kimi K2 Thinking has 1T parameters, while GLM-4.7-Flash has 30B. Context windows are 256k (Kimi K2 Thinking) vs 200k (GLM-4.7-Flash).

On SWE-Bench Verified, Kimi K2 Thinking leads at 71.3% vs GLM-4.7-Flash at 59.2%. On LiveCodeBench, Kimi K2 Thinking leads at 83.1% vs GLM-4.7-Flash at 64%. On BrowseComp, Kimi K2 Thinking leads at 60.2% vs GLM-4.7-Flash at 42.8%. On AIME 2025, Kimi K2 Thinking leads at 94.5% vs GLM-4.7-Flash at 91.6%.