GLM 5.3 vs Kimi K3

GLM 5.3
$0.85 / task
Wins 1 of 5 benchmarks
Kimi K3
$2.00 / task
Wins 3 of 5 benchmarks
Different efforts: low · max · Compare all at max
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
GLM 5.336.6
Kimi K346.9
Terminal-Bench 4.0 Terminal-Bench 4.0 · source setting 2: Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite
GLM 5.334.8
Kimi K312.6
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
GLM 5.352
Kimi K358.3
Arena Text · creative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool
GLM 5.3 · max1,453
Kimi K31,461
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos
GLM 5.3—
Kimi K380.5

Specs

SpecGLM 5.3Kimi K3
Input, per 1M!$1.40$3
Output, per 1M!$4.40$15
Cost per task?$0.85$2.00
Speed, tok/s126112
Context1M1M
WeightsOpenOpen
ReleasedAug 2026Jul 2026
Memory (4-bit)452 GB1,668 GB
Licenceglm-5.3kimi-k3
Commercial useRestrictedRestricted
See images—✓
Read PDFs—Partial
Hear audio——
Watch video—✓
Call tools✓✓
Force a tool call—✓
Return strict JSON✓✓
Search the web?Partial
Run code??
Use a computer??

What breaks when you switch

GLM 5.3 → Kimi K3
Memory (4-bit): 452 GB → 1,668 GB
Sampling parameters are fixed›

temperature, top_p, n and penalty values are fixed; omit them from requests.

No public image URLs›

Vision input rejects public image URLs; use base64 or ms:// file references in a content array.

Web search is being updated›

Web search is not recommended for near-term production workflows; avoid depending on it in production.

Thinking mode is always enabled›

K3 always has thinking enabled; control effort with the top-level reasoning_effort field using low, high or max.

Kimi K3 → GLM 5.3
No image input
Reasoning cannot be disabled›

Disabling reasoning is unsupported; keep reasoning enabled.

tool_choice auto · Z.ai API›

The first-party chat reference exposes only auto for tool_choice.

Complete unmodified reasoning_content›

Preserved thinking defaults on for Coding Plan and off for the standard API; when enabled, return complete unmodified reasoning_content in its original order to protect quality and cache hits.

Only one stop word›

Only one stop word is supported, so do not rely on a list of alternative stop sequences.