Compare · ModelsLive · 2 picked · head to head
Gemini 3 Flash Preview vs GLM 4.6
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Gemini 3 Flash Preview wins on 6/6 benchmarks
Gemini 3 Flash Preview wins 6 of 6 shared benchmarks. Leads in agentic · arena · math.
Category leads
agentic·Gemini 3 Flash Previewarena·Gemini 3 Flash Previewmath·Gemini 3 Flash Previewcoding·Gemini 3 Flash Preview
Hype vs Reality
Attention vs performance
Gemini 3 Flash Preview
#125 by perf·#5 by attention
GLM 4.6
#117 by perf·#3 by attention
Best value
GLM 4.6
1.6x better value than Gemini 3 Flash Preview
Gemini 3 Flash Preview
28.3 pts/$
$1.75/M
GLM 4.6
46.6 pts/$
$1.09/M
Vendor risk
Who is behind the model
Google DeepMind
$4.20T·Tier 1
z-ai
private · undisclosed
Head to head
6 benchmarks · 2 models
Gemini 3 Flash PreviewGLM 4.6
APEX-Agents
Gemini 3 Flash Preview leads by +21.0
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
Gemini 3 Flash Preview
24.0
GLM 4.6
3.0
Chatbot Arena Elo · Coding
Gemini 3 Flash Preview leads by +99.5
Gemini 3 Flash Preview
1438.9
GLM 4.6
1339.4
Chatbot Arena Elo · Overall
Gemini 3 Flash Preview leads by +48.6
Gemini 3 Flash Preview
1472.7
GLM 4.6
1424.1
FrontierMath-2025-02-28-Private
Gemini 3 Flash Preview leads by +31.8
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
Gemini 3 Flash Preview
35.6
GLM 4.6
3.8
FrontierMath-Tier-4-2025-07-01-Private
Gemini 3 Flash Preview leads by +2.0
FrontierMath Tier 4 (Jul 2025) · the most challenging tier of frontier mathematics, containing problems that push the absolute limits of AI mathematical reasoning.
Gemini 3 Flash Preview
4.2
GLM 4.6
2.1
Terminal Bench
Gemini 3 Flash Preview leads by +39.8
Terminal-Bench 2.0 · evaluates AI agents on real terminal-based coding tasks · writing scripts, debugging, running tests, and managing projects entirely through command-line interaction. Tests both code quality and terminal fluency. Claude Opus 4.7 scores 69.4%, demonstrating significant agentic terminal competence.
Gemini 3 Flash Preview
64.3
GLM 4.6
24.5
Full benchmark table
| Benchmark | Gemini 3 Flash Preview | GLM 4.6 |
|---|---|---|
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | 24.0 | 3.0 |
Chatbot Arena Elo · Coding | 1438.9 | 1339.4 |
Chatbot Arena Elo · Overall | 1472.7 | 1424.1 |
FrontierMath-2025-02-28-Private FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning. | 35.6 | 3.8 |
FrontierMath-Tier-4-2025-07-01-Private FrontierMath Tier 4 (Jul 2025) · the most challenging tier of frontier mathematics, containing problems that push the absolute limits of AI mathematical reasoning. | 4.2 | 2.1 |
Terminal Bench Terminal-Bench 2.0 · evaluates AI agents on real terminal-based coding tasks · writing scripts, debugging, running tests, and managing projects entirely through command-line interaction. Tests both code quality and terminal fluency. Claude Opus 4.7 scores 69.4%, demonstrating significant agentic terminal competence. | 64.3 | 24.5 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.50 | $3.00 | 1.0M tokens (~524 books) | $11.25 | |
| $0.43 | $1.75 | 205K tokens (~102 books) | $7.60 |