Compare · ModelsLive · 3 picked · head to head
GLM 5 vs GPT-5.1-Codex-Max vs Qwen3.6 Plus
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Qwen3.6 Plus wins on 7/16 benchmarks
Qwen3.6 Plus wins 7 of 16 shared benchmarks. Leads in reasoning · math · arena.
Category leads
coding·GPT-5.1-Codex-Maxreasoning·Qwen3.6 Pluslanguage·GPT-5.1-Codex-Maxmath·Qwen3.6 Plusknowledge·GPT-5.1-Codex-Maxarena·Qwen3.6 Plus
Hype vs Reality
Attention vs performance
GLM 5
#79 by perf·#27 by attention
GPT-5.1-Codex-Max
#16 by perf·no signal
Qwen3.6 Plus
#55 by perf·no signal
Best value
Qwen3.6 Plus
1.2x better value than GLM 5
GLM 5
44.5 pts/$
$1.26/M
GPT-5.1-Codex-Max
12.8 pts/$
$5.63/M
Qwen3.6 Plus
52.7 pts/$
$1.14/M
Vendor risk
Who is behind the model
z-ai
private · undisclosed
OpenAI
$840.0B·Tier 1
Alibaba (Qwen)
$293.0B·Tier 1
Head to head
16 benchmarks · 3 models
GLM 5GPT-5.1-Codex-MaxQwen3.6 Plus
LiveBench · Agentic Coding
GPT-5.1-Codex-Max leads by +1.7
GLM 5
55.0
GPT-5.1-Codex-Max
56.7
Qwen3.6 Plus
55.0
LiveBench · Coding
GPT-5.1-Codex-Max leads by +3.2
GLM 5
73.6
GPT-5.1-Codex-Max
81.4
Qwen3.6 Plus
78.2
LiveBench · Data Analysis
Qwen3.6 Plus leads by +2.0
GLM 5
67.9
GPT-5.1-Codex-Max
54.9
Qwen3.6 Plus
69.9
LiveBench · If
GPT-5.1-Codex-Max leads by +8.8
GLM 5
55.3
GPT-5.1-Codex-Max
67.1
Qwen3.6 Plus
58.3
LiveBench · Language
GLM 5 leads by +2.1
GLM 5
77.5
GPT-5.1-Codex-Max
75.4
Qwen3.6 Plus
75.0
LiveBench · Mathematics
Qwen3.6 Plus leads by +0.1
GLM 5
83.5
GPT-5.1-Codex-Max
83.7
Qwen3.6 Plus
83.7
LiveBench · Overall
GPT-5.1-Codex-Max leads by +1.1
GLM 5
68.8
GPT-5.1-Codex-Max
72.0
Qwen3.6 Plus
70.8
LiveBench · Reasoning
GPT-5.1-Codex-Max leads by +8.7
GLM 5
69.1
GPT-5.1-Codex-Max
84.6
Qwen3.6 Plus
75.8
Chatbot Arena Elo · Coding
Qwen3.6 Plus leads by +27.4
GLM 5
1434.5
Qwen3.6 Plus
1461.9
Chatbot Arena Elo · Overall
GLM 5 leads by +13.0
GLM 5
1457.3
Qwen3.6 Plus
1444.2
Chess Puzzles
Qwen3.6 Plus leads by +6.2
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
GLM 5
10.0
Qwen3.6 Plus
16.2
FrontierMath-2025-02-28-Private
Qwen3.6 Plus leads by +9.8
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
GLM 5
16.4
Qwen3.6 Plus
26.2
FrontierMath-Tier-4-2025-07-01-Private
Qwen3.6 Plus leads by +6.2
FrontierMath Tier 4 (Jul 2025) · the most challenging tier of frontier mathematics, containing problems that push the absolute limits of AI mathematical reasoning.
GLM 5
2.1
Qwen3.6 Plus
8.3
GPQA diamond
GLM 5 leads by +0.6
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
GLM 5
83.8
Qwen3.6 Plus
83.2
OTIS Mock AIME 2024-2025
Qwen3.6 Plus leads by +10.6
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
GLM 5
80.0
Qwen3.6 Plus
90.5
SWE-Bench verified
GLM 5 leads by +14.2
SWE-bench Verified · 500 human-validated tasks from 12 real Python repositories (Django, Flask, scikit-learn, sympy, and others). Each task requires the model to produce a git patch that resolves a real GitHub issue and passes the test suite. The verified subset eliminates ambiguous tasks from the original SWE-bench. Claude Mythos Preview leads at 93.9%, crossing 90% for the first time in 2026. Opus 4.6 scores 80.8%. The benchmark remains the most-cited evaluation for code-generation capability.
GLM 5
72.1
Qwen3.6 Plus
57.9
Full benchmark table
| Benchmark | GLM 5 | GPT-5.1-Codex-Max | Qwen3.6 Plus |
|---|---|---|---|
LiveBench · Agentic Coding | 55.0 | 56.7 | 55.0 |
LiveBench · Coding | 73.6 | 81.4 | 78.2 |
LiveBench · Data Analysis | 67.9 | 54.9 | 69.9 |
LiveBench · If | 55.3 | 67.1 | 58.3 |
LiveBench · Language | 77.5 | 75.4 | 75.0 |
LiveBench · Mathematics | 83.5 | 83.7 | 83.7 |
LiveBench · Overall | 68.8 | 72.0 | 70.8 |
LiveBench · Reasoning | 69.1 | 84.6 | 75.8 |
Chatbot Arena Elo · Coding | 1434.5 | — | 1461.9 |
Chatbot Arena Elo · Overall | 1457.3 | — | 1444.2 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 10.0 | — | 16.2 |
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. | 16.4 | — | 26.2 |
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. | 2.1 | — | 8.3 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 83.8 | — | 83.2 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 80.0 | — | 90.5 |
SWE-Bench verified SWE-bench Verified · 500 human-validated tasks from 12 real Python repositories (Django, Flask, scikit-learn, sympy, and others). Each task requires the model to produce a git patch that resolves a real GitHub issue and passes the test suite. The verified subset eliminates ambiguous tasks from the original SWE-bench. Claude Mythos Preview leads at 93.9%, crossing 90% for the first time in 2026. Opus 4.6 scores 80.8%. The benchmark remains the most-cited evaluation for code-generation capability. | 72.1 | — | 57.9 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.60 | $1.92 | 203K tokens (~101 books) | $9.30 | |
| $1.25 | $10.00 | 400K tokens (~200 books) | $34.38 | |
| $0.33 | $1.95 | 1.0M tokens (~500 books) | $7.31 |
People also compared