Compare · ModelsLive · 2 picked · head to head
Gemini 3.1 Pro Preview vs GPT-5.3-Codex
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Gemini 3.1 Pro Preview wins on 13/17 benchmarks
Gemini 3.1 Pro Preview wins 13 of 17 shared benchmarks. Leads in speed · agentic · general.
Category leads
speed·Gemini 3.1 Pro Previewagentic·Gemini 3.1 Pro Previewgeneral·Gemini 3.1 Pro Previewknowledge·Gemini 3.1 Pro Previewcoding·Gemini 3.1 Pro Preview
Head to head · from the data · data as of 2026-11-17
- Benchmarks · Gemini 3.1 Pro Preview leads on 13 of 17 shared benchmarks (Artificial Analysis · Coding Index, Artificial Analysis · CritPt, Artificial Analysis · GPQA Diamond, Artificial Analysis · Humanity's Last Exam, 9 more), GPT-5.3-Codex on 4 (Artificial Analysis · Agentic Index, Artificial Analysis · Long Context Reasoning, Artificial Analysis · Quality Index, WeirdML)
- Price · Gemini 3.1 Pro Preview input $2.00/M vs GPT-5.3-Codex $1.75/M, GPT-5.3-Codex is 13% cheaper (OpenRouter list price, as of 2026-11-17)
- Cheapest provider · Gemini 3.1 Pro Preview at Google and 1 other provider $1.00/M input (of 6 offers) as of 2026-11-17 · GPT-5.3-Codex at Azure and 1 other provider $1.75/M input (of 3 offers) as of 2026-11-17
- Context · Gemini 3.1 Pro Preview 1.0M tokens vs GPT-5.3-Codex 400K tokens
- Released · Gemini 3.1 Pro Preview Feb 19, 2026 vs GPT-5.3-Codex Feb 24, 2026 · the same month
- Price history · Gemini 3.1 Pro Preview OpenRouter input list price had no move of 5% or more in the 90 days to 2026-11-17 · GPT-5.3-Codex OpenRouter input list price had no move of 5% or more in the 90 days to 2026-11-17
Sources · scores from public leaderboards, list prices and provider offers from OpenRouter, price moves from the BenchGecko changelog (5% or more, held in the next daily check), Gecko tests run by BenchGecko.
Hype vs Reality
Attention vs performance
Gemini 3.1 Pro Preview
#143 by perf·#9 by attention
GPT-5.3-Codex
#86 by perf·#11 by attention
Best value
GPT-5.3-Codex
1.0x better value than Gemini 3.1 Pro Preview
Gemini 3.1 Pro Preview
7.0 pts/$
$7.00/M
GPT-5.3-Codex
7.0 pts/$
$7.88/M
Vendor risk
Who is behind the model
Google DeepMind
$3.49T·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
17 benchmarks · 2 models
Gemini 3.1 Pro PreviewGPT-5.3-Codex
Artificial Analysis · Agentic Index
GPT-5.3-Codex leads by +40.8
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?"
Gemini 3.1 Pro Preview
21.4
GPT-5.3-Codex
62.2
Artificial Analysis · Coding Index
Gemini 3.1 Pro Preview leads by +15.7
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.
Gemini 3.1 Pro Preview
68.8
GPT-5.3-Codex
53.1
Artificial Analysis · CritPt
Gemini 3.1 Pro Preview leads by +0.8
Gemini 3.1 Pro Preview
17.7
GPT-5.3-Codex
16.9
Artificial Analysis · GPQA Diamond
Gemini 3.1 Pro Preview leads by +2.6
Gemini 3.1 Pro Preview
94.1
GPT-5.3-Codex
91.5
Artificial Analysis · Humanity's Last Exam
Gemini 3.1 Pro Preview leads by +4.5
Gemini 3.1 Pro Preview
47.0
GPT-5.3-Codex
42.5
Artificial Analysis · IFBench
Gemini 3.1 Pro Preview leads by +1.7
Gemini 3.1 Pro Preview
77.1
GPT-5.3-Codex
75.4
Artificial Analysis · Long Context Reasoning
GPT-5.3-Codex leads by +1.3
Gemini 3.1 Pro Preview
82.0
GPT-5.3-Codex
83.3
Artificial Analysis · MMMU Pro
Gemini 3.1 Pro Preview leads by +3.9
Gemini 3.1 Pro Preview
82.4
GPT-5.3-Codex
78.5
Artificial Analysis · Quality Index
GPT-5.3-Codex leads by +2.8
Gemini 3.1 Pro Preview
29.7
GPT-5.3-Codex
32.5
Artificial Analysis · tau2-Bench Telecom
Gemini 3.1 Pro Preview leads by +9.6
Gemini 3.1 Pro Preview
95.6
GPT-5.3-Codex
86.0
Artificial Analysis · Terminal-Bench Hard
Gemini 3.1 Pro Preview leads by +0.8
Gemini 3.1 Pro Preview
53.8
GPT-5.3-Codex
53.0
APEX-Agents
Gemini 3.1 Pro Preview leads by +3.4
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
Gemini 3.1 Pro Preview
35.1
GPT-5.3-Codex
31.7
Metr Time Horizons
Gemini 3.1 Pro Preview leads by +2.5
Gemini 3.1 Pro Preview
77.0
GPT-5.3-Codex
74.5
PostTrainBench
Gemini 3.1 Pro Preview leads by +0.3
Gemini 3.1 Pro Preview
18.1
GPT-5.3-Codex
17.8
SWE-Bench verified
Gemini 3.1 Pro Preview leads by +0.8
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. The benchmark remains the most-cited evaluation for code-generation capability.
Gemini 3.1 Pro Preview
75.6
GPT-5.3-Codex
74.8
Terminal Bench
Gemini 3.1 Pro Preview leads by +2.9
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.
Gemini 3.1 Pro Preview
80.2
GPT-5.3-Codex
77.3
WeirdML
GPT-5.3-Codex leads by +7.2
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Gemini 3.1 Pro Preview
72.1
GPT-5.3-Codex
79.3
Full benchmark table
| Benchmark | Gemini 3.1 Pro Preview | GPT-5.3-Codex |
|---|---|---|
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?" | 21.4 | 62.2 |
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. | 68.8 | 53.1 |
Artificial Analysis · CritPt | 17.7 | 16.9 |
Artificial Analysis · GPQA Diamond | 94.1 | 91.5 |
Artificial Analysis · Humanity's Last Exam | 47.0 | 42.5 |
Artificial Analysis · IFBench | 77.1 | 75.4 |
Artificial Analysis · Long Context Reasoning | 82.0 | 83.3 |
Artificial Analysis · MMMU Pro | 82.4 | 78.5 |
Artificial Analysis · Quality Index | 29.7 | 32.5 |
Artificial Analysis · tau2-Bench Telecom | 95.6 | 86.0 |
Artificial Analysis · Terminal-Bench Hard | 53.8 | 53.0 |
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | 35.1 | 31.7 |
Metr Time Horizons | 77.0 | 74.5 |
PostTrainBench | 18.1 | 17.8 |
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. The benchmark remains the most-cited evaluation for code-generation capability. | 75.6 | 74.8 |
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. | 80.2 | 77.3 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 72.1 | 79.3 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $2.00 | $12.00 | 1.0M tokens (~524 books) | $45.00 | |
| $1.75 | $14.00 | 400K tokens (~200 books) | $48.13 |
Gemini 3.1 Pro Preview vs GPT-5.3-Codex · questions answered from the data
Gemini 3.1 Pro Preview leads on 13 of 17 shared benchmarks (Artificial Analysis · Coding Index, Artificial Analysis · CritPt, Artificial Analysis · GPQA Diamond, Artificial Analysis · Humanity's Last Exam, 9 more), GPT-5.3-Codex on 4 (Artificial Analysis · Agentic Index, Artificial Analysis · Long Context Reasoning, Artificial Analysis · Quality Index, WeirdML)