Compare/Kimi K2 0905 vs GLM-4.6 (Non-reasoning)

Kimi K2 0905vsGLM-4.6 (Non-reasoning)

Side-by-side comparison of pricing, 12 benchmarks, and generation speed.

Kimi

Kimi K2 0905

Input
$0.8/M
Output
$2.25/M
Speed
63 tok/s
TTFT
0.78s
Z AI

GLM-4.6 (Non-reasoning)

Input
$0.6/M
Output
$2.2/M
Speed
86 tok/s
TTFT
1.68s

Winner by Category

Cheaper
GLM-4.6 (Non-reasoning)
Faster (tok/s)
GLM-4.6 (Non-reasoning)
Lower Latency
Kimi K2 0905
Benchmarks (8-3)
Kimi K2 0905

Pricing Comparison

MetricKimi K2 0905GLM-4.6 (Non-reasoning)
Input ($/M tokens)$0.8$0.6
Output ($/M tokens)$2.25$2.2
Cost for 1M input + 100K output tokens:
Kimi K2 0905$1.03
GLM-4.6 (Non-reasoning)$0.82

Speed Comparison

Output Speed (tokens/s) — higher is better
Kimi K2 0905
63 tok/s
GLM-4.6 (Non-reasoning)
86 tok/s
Time to First Token (seconds) — lower is better
Kimi K2 0905
0.78s
GLM-4.6 (Non-reasoning)
1.68s

Benchmark Comparison

Data from Artificial Analysis API — 12 benchmarks

Intelligence Index
30.930.2
Coding Index
25.930.2
Math Index
57.344.3
GPQA Diamond
76.7%63.2%
MMLU-Pro
81.9%78.4%
LiveCodeBench
61.0%56.1%
AIME 2025
57.3%44.3%
MATH-500
Humanity's Last Exam
6.3%5.2%
SciCode
30.7%33.1%
IFBench
41.7%36.7%
TerminalBench
23.5%28.8%
Kimi K2 09058 wins
3 winsGLM-4.6 (Non-reasoning)

Frequently Asked Questions

Which is cheaper, Kimi K2 0905 or GLM-4.6 (Non-reasoning)?

GLM-4.6 (Non-reasoning) is cheaper overall. Its blended price (3:1 input/output ratio) is $1.00/M tokens vs $1.14/M for Kimi K2 0905.

Which model performs better on benchmarks?

Kimi K2 0905 wins 8 out of 12 benchmarks compared to 3 for GLM-4.6 (Non-reasoning). See the detailed benchmark chart above for per-category results.

Which is faster for real-time applications?

GLM-4.6 (Non-reasoning) generates tokens faster at 86 tok/s vs 63 tok/s. Kimi K2 0905 also has lower time-to-first-token (0.78s vs 1.68s).

When should I use Kimi K2 0905 vs GLM-4.6 (Non-reasoning)?

Choose based on your priorities: GLM-4.6 (Non-reasoning) for lower cost, Kimi K2 0905 for stronger benchmark performance, and GLM-4.6 (Non-reasoning) for faster generation. For latency-sensitive apps, check the TTFT comparison above.