Compare / head-to-head
Gemini 3.5 Flashvs
GPT-5.3-Codex
GPT-5.3-Codex leads 3 of 5 shared benchmarks. Gemini 3.5 Flash is 1.5x cheaper per token. Gemini 3.5 Flash has the larger context window (1.0M tokens).
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
2 – 3
GPT-5.3-Codex leads
Cheaper per token
Gemini 3.5 Flash
1.5x cheaper, input + output
Larger context
Gemini 3.5 Flash
1.0M tokens
Providers
2 providers
Google · OpenAI
Gemini 3.5 Flash
Google · released 2026-05-19
textvisionaudio
GPT-5.3-Codex
OpenAI · released 2026-02-05
textvision
Quality
Benchmark matrix
5 shared · 31 only Gemini 3.5 Flash · 4 only GPT-5.3-Codex| Benchmark | Gemini 3.5 Flash | GPT-5.3-Codex | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 5 | ||||
| OSWorld-Verified | 78.4% ↗ | 64.7% ↗ | +13.7 pt | Gemini 3.5 Flash |
| SWE-Bench Pro (Public) | 55.1% ↗ | 56.8% ↗ | -1.7 pt | GPT-5.3-Codex |
| SWE-bench Verified (Epoch AI run) | 79.3% ↗epoch run | 74.8% ↗epoch run | +4.5 pt | Gemini 3.5 Flash |
| Vending-Bench 2 (Andon Labs) | 5396.42 ↗ | 5940.12 ↗ | -543.7 | GPT-5.3-Codex |
| WeirdML (Håvard Tveit Ihle) | 62.6% ↗ | 79.3% ↗ | -16.7 pt | GPT-5.3-Codex |
| Only Gemini 3.5 Flash reports · 31 | ||||
| GPQA Diamond | 92.8% ↗epoch run | not reported | — | — |
| AIME 2026 | 95% ↗matharena ⚠ | not reported | — | — |
| LMArena Elo | 1477.4 ↗ | not reported | — | — |
| APEX-Agents (Mercor) | 27.5% ↗ | not reported | — | — |
| ARC-AGI-2 | 72.1% ↗ | not reported | — | — |
| Chess Puzzles (Epoch AI run) | 50% ↗epoch run | not reported | — | — |
| DeepSWE v1.1 (Datacurve) | 36.1% ↗ | not reported | — | — |
| EBR-bench (Epoch AI run) | 4.8% ↗epoch run | not reported | — | — |
| FrontierMath Tier 4 v2 (Epoch AI run) | 26.8% ↗epoch run | not reported | — | — |
| FrontierMath Tiers 1-3 v2 (Epoch AI run) | 62.8% ↗epoch run | not reported | — | — |
| GDPVal-AA | 1656 ↗ | not reported | — | — |
| Humanity's Last Exam (full set, text + MM) | 40.2% ↗ | not reported | — | — |
| LiveBench Agentic Coding (LiveBench) | 49% ↗ | not reported | — | — |
| LiveBench Coding (LiveBench) | 78.2% ↗ | not reported | — | — |
| LiveBench Data Analysis (LiveBench) | 64.9% ↗ | not reported | — | — |
| LiveBench Instruction Following (LiveBench) | 75.6% ↗ | not reported | — | — |
| LiveBench Language (LiveBench) | 84.6% ↗ | not reported | — | — |
| LiveBench Mathematics (LiveBench) | 88.2% ↗ | not reported | — | — |
| LiveBench Reasoning (LiveBench) | 82% ↗ | not reported | — | — |
| LMArena Vision (LMArena) | 1284.4 ↗ | not reported | — | — |
| LMArena WebDev (LMArena) | 1499 ↗ | not reported | — | — |
| MCP Atlas | 83.6% ↗ | not reported | — | — |
| MMMU-Pro | 83.6% ↗ | not reported | — | — |
| Mystery Game Puzzles (Epoch AI run) | 32% ↗epoch run | not reported | — | — |
| OTIS Mock AIME 2024-2025 (Epoch AI run) | 95.6% ↗epoch run | not reported | — | — |
| SAGE (Vals AI) | 49.9% ↗ | not reported | — | — |
| SimpleBench (SimpleBench) | 76.7% ↗ | not reported | — | — |
| SimpleQA Verified | 66.2% ↗epoch run | not reported | — | — |
| Terminal-bench 2.1 | 76.2% ↗ | not reported | — | — |
| Terminal-Bench 4.0 (Vals AI) | 6.1% ↗ | not reported | — | — |
| Toolathlon-Verified (HKUST) | 67.3% ↗ | not reported | — | — |
| Only GPT-5.3-Codex reports · 4 | ||||
| Cybersecurity Capture The Flag Challenges | not reported | 77.6% ↗ | — | — |
| GDPval (wins or ties) | not reported | 70.9% ↗ | — | — |
| SWE-Lancer IC Diamond | not reported | 81.4% ↗ | — | — |
| Terminal-Bench 2.0 | not reported | 77.3% ↗ | — | — |
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is Gemini 3.5 Flash minus GPT-5.3-Codex in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing
Side by side
Official pricing pages and model cards| Spec | Gemini 3.5 Flash | GPT-5.3-Codex | Edge |
|---|---|---|---|
| Context window | 1.0M tokens | 400K tokens | Gemini 3.5 Flash |
| Max output | 66K tokens | 128K tokens | GPT-5.3-Codex |
| Input price / 1M | $1.5 ↗ | $1.75 ↗ | Gemini 3.5 Flash |
| Output price / 1M | $9 ↗ | $14 ↗ | Gemini 3.5 Flash |
| Cached input / 1M | $0.15 ↗ | $0.175 ↗ | Gemini 3.5 Flash |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | $10.5 | $15.75 | Gemini 3.5 Flash |
| Modalities | text · vision · audio | text · vision | Gemini 3.5 Flash |
| Released | 2026-05-19 | 2026-02-05 | — |
| Cited benchmark scores | 42 | 9 | — |
Reliability
Provider status
Live from /statusMore matchups
Gemini 3.5 Flash vs …
Models sharing the most benchmarksMore matchups
GPT-5.3-Codex vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run Gemini 3.5 Flash and GPT-5.3-Codex on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.