Compare · ModelsLive · 3 picked · head to head
GPT-5.4 vs Claude Opus 4.6 (Fast) vs GLM 5.1
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GPT-5.4 wins on 11/13 benchmarks
GPT-5.4 wins 11 of 13 shared benchmarks. Leads in speed · knowledge · math.
Category leads
speed·GPT-5.4arena·Claude Opus 4.6 (Fast)knowledge·GPT-5.4math·GPT-5.4coding·GPT-5.4
Hype vs Reality
Attention vs performance
GPT-5.4
#61 by perf·no signal
Claude Opus 4.6 (Fast)
#150 by perf·no signal
GLM 5.1
#56 by perf·no signal
Best value
GLM 5.1
4.4x better value than GPT-5.4
GPT-5.4
6.7 pts/$
$8.75/M
Claude Opus 4.6 (Fast)
0.5 pts/$
$90.00/M
GLM 5.1
29.9 pts/$
$2.00/M
Vendor risk
Who is behind the model
OpenAI
$840.0B·Tier 1
Anthropic
$380.0B·Tier 1
z-ai
private · undisclosed
Head to head
13 benchmarks · 3 models
GPT-5.4Claude Opus 4.6 (Fast)GLM 5.1
Artificial Analysis · Agentic Index
GPT-5.4 leads by +1.8
Artificial Analysis Agentic Index · a composite score measuring how well a model performs in agentic workflows · multi-step tool use, planning, error recovery, and autonomous task completion. Aggregates results from multiple agentic benchmarks including SWE-bench, tool-use tests, and planning evaluations. The canonical single-number metric for "how good is this model as an agent?"
GPT-5.4
69.4
Claude Opus 4.6 (Fast)
67.6
GLM 5.1
29.9
Artificial Analysis · Coding Index
GPT-5.4 leads by +1.5
Artificial Analysis Coding Index · a composite score that aggregates performance across multiple coding benchmarks into a single index. Tracks code generation quality, debugging ability, multi-language competence, and real-world software engineering tasks. Used by Artificial Analysis to rank model coding capability in a normalized, comparable format. Useful for developers choosing between models for coding-heavy workloads.
GPT-5.4
57.3
Claude Opus 4.6 (Fast)
48.1
GLM 5.1
55.8
Artificial Analysis · Quality Index
GPT-5.4 leads by +4.2
GPT-5.4
57.2
Claude Opus 4.6 (Fast)
53.0
GLM 5.1
40.2
Chatbot Arena Elo · Coding
Claude Opus 4.6 (Fast) leads by +13.0
GPT-5.4
1411.1
Claude Opus 4.6 (Fast)
1542.2
GLM 5.1
1529.2
Chatbot Arena Elo · Overall
Claude Opus 4.6 (Fast) leads by +28.4
GPT-5.4
1467.7
Claude Opus 4.6 (Fast)
1503.7
GLM 5.1
1475.3
Chess Puzzles
GPT-5.4 leads by +26.0
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
GPT-5.4
44.0
GLM 5.1
18.0
FrontierMath-2025-02-28-Private
GPT-5.4 leads by +14.1
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
GPT-5.4
47.6
GLM 5.1
33.5
FrontierMath-Tier-4-2025-07-01-Private
GPT-5.4 leads by +14.6
FrontierMath Tier 4 (Jul 2025) · the most challenging tier of frontier mathematics, containing problems that push the absolute limits of AI mathematical reasoning.
GPT-5.4
27.1
GLM 5.1
12.5
GPQA diamond
GPT-5.4 leads by +10.4
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
GPT-5.4
91.1
GLM 5.1
80.6
OTIS Mock AIME 2024-2025
GPT-5.4 leads by +3.1
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
GPT-5.4
95.3
GLM 5.1
92.2
SimpleQA Verified
GPT-5.4 leads by +7.5
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
GPT-5.4
44.8
GLM 5.1
37.3
SWE-Bench verified
GPT-5.4 leads by +2.7
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.
GPT-5.4
76.9
GLM 5.1
74.2
WeirdML
GPT-5.4 leads by +20.6
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
GPT-5.4
77.7
GLM 5.1
57.1
Full benchmark table
| Benchmark | GPT-5.4 | Claude Opus 4.6 (Fast) | GLM 5.1 |
|---|---|---|---|
Artificial Analysis · Agentic Index Artificial Analysis Agentic Index · a composite score measuring how well a model performs in agentic workflows · multi-step tool use, planning, error recovery, and autonomous task completion. Aggregates results from multiple agentic benchmarks including SWE-bench, tool-use tests, and planning evaluations. The canonical single-number metric for "how good is this model as an agent?" | 69.4 | 67.6 | 29.9 |
Artificial Analysis · Coding Index Artificial Analysis Coding Index · a composite score that aggregates performance across multiple coding benchmarks into a single index. Tracks code generation quality, debugging ability, multi-language competence, and real-world software engineering tasks. Used by Artificial Analysis to rank model coding capability in a normalized, comparable format. Useful for developers choosing between models for coding-heavy workloads. | 57.3 | 48.1 | 55.8 |
Artificial Analysis · Quality Index | 57.2 | 53.0 | 40.2 |
Chatbot Arena Elo · Coding | 1411.1 | 1542.2 | 1529.2 |
Chatbot Arena Elo · Overall | 1467.7 | 1503.7 | 1475.3 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 44.0 | — | 18.0 |
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. | 47.6 | — | 33.5 |
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. | 27.1 | — | 12.5 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 91.1 | — | 80.6 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 95.3 | — | 92.2 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 44.8 | — | 37.3 |
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. | 76.9 | — | 74.2 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 77.7 | — | 57.1 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $2.50 | $15.00 | 1.1M tokens (~525 books) | $56.25 | |
| $30.00 | $150.00 | 1.0M tokens (~500 books) | $600.00 | |
| $0.97 | $3.04 | 203K tokens (~101 books) | $14.83 |