Model comparison
Gemini 3.8 Flash vs o3
Gemini 3.8 Flash is the stronger model overall, scoring 61.8 to 47.5 on the Noometry Index.
Last verified . 34 shared benchmarks.
GoogleSummary
- They share 34 benchmarks with published results for both. Gemini 3.8 Flash scores higher in 8 categories and o3 in 2 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3.8 Flash leads 76.9 to 32.0.
- The biggest single-benchmark swing is ARC-AGI-2: 89.2% for Gemini 3.8 Flash and 6.5% for o3.
- Gemini 3.8 Flash is cheaper at $0.75 / $3.75 per million input/output tokens, against $2 / $8 for o3.
- Gemini 3.8 Flash accepts more context: 1.05M tokens versus 200K.
Side by side
| Gemini 3.8 Flash | o3 | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 61.8 | 47.5 |
| Released | 2026-09-02 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 200K |
| Max output | 66K | 100K |
| Input $ / M tokens | $0.75 | $2 |
| Output $ / M tokens | $3.75 | $8 |
| Results tracked | 50 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3.8 Flash leads
Gemini 3.8 Flash: 59.2 (#15), o3: 46.8 (#64)
| Benchmark | Gemini 3.8 Flash | o3 |
|---|---|---|
| WeirdML | 84.8% | 52.4% |
| LMArena Coding | 1510 | 1408 |
| ALE-Bench | 1,270 | 933.55 |
| SWE-bench Verified | — | 62.3% |
| DeepSWE | 73.8% | — |
| FrontierCode | 41.2% | — |
| SWE-bench Verified (bash only) | — | 58.4% |
| Aider Polyglot | — | 81.3% |
| CursorBench | 39.6% | — |
| LMArena WebDev | 1584 | — |
| FrontierSWE | 19.6% | — |
| SciCode | 56.6% | — |
| GSO | — | 8.8% |
| CadEval | — | 74% |
Agentic & Tool Use Gemini 3.8 Flash leads
Gemini 3.8 Flash: 41.8 (#21), o3: 34.5 (#44)
| Benchmark | Gemini 3.8 Flash | o3 |
|---|---|---|
| APEX-Agents | 64.3% | — |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 30.8% |
| Remote Labor Index | 5.8% | — |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| GDP.pdf | 23.4% | — |
| LMArena Search | — | 1144 |
| METR Time Horizons | — | 65.4% |
| Vending-Bench 2 | 5,094 | — |
Reasoning Gemini 3.8 Flash leads
Gemini 3.8 Flash: 76.9 (#5), o3: 32.0 (#78)
| Benchmark | Gemini 3.8 Flash | o3 |
|---|---|---|
| ARC-AGI-2 | 89.2% | 6.5% |
| ARC-AGI-1 | 98.5% | 60.8% |
| CritPt | 18.3% | 1.4% |
| Chess Puzzles | 61% | 38% |
| LMArena Hard Prompts | 1508 | 1402 |
| Mystery Game Puzzles | 47% | 29% |
| DTBench | 95.7% | 84.8% |
| LMCA | 52.9% | 39.7% |
| Epoch Capabilities Index | 156.71 | 146.86 |
| SimpleBench | — | 53.1% |
| Kagi LLM Benchmark | — | 67.6% |
| NYT Connections (extended) | 97.4% | — |
| EnigmaEval | — | 13.1% |
| Surface Evolver Bench | 76.9% | — |
| ForecastBench | — | 62.5 |
Math Gemini 3.8 Flash leads
Gemini 3.8 Flash: 65.3 (#28), o3: 50.2 (#58)
| Benchmark | Gemini 3.8 Flash | o3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 68.4% | 33.3% |
| OTIS Mock AIME 2024-2025 | 98.9% | 84.4% |
| LMArena Math | 1528 | 1426 |
| FrontierMath Tier 4 | 22% | — |
| ProofBench | 48% | — |
| Omni-MATH | — | 71.4% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 18.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge Gemini 3.8 Flash leads
Gemini 3.8 Flash: 74.8 (#2), o3: 54.6 (#52)
| Benchmark | Gemini 3.8 Flash | o3 |
|---|---|---|
| GPQA Diamond | 95.4% | 81.8% |
| Humanity's Last Exam | 44.5% | 20.3% |
| SimpleQA Verified | 69.7% | 49.4% |
| LMArena Expert | 1524 | 1402 |
| MMLU-Pro | — | 85.9% |
| Confabulations | — | 14.4% |
| GPQA (HELM) | — | 75.3% |
Multimodal Too close to call
Gemini 3.8 Flash: 40.7 (#45), o3: 41.4 (#36)
| Benchmark | Gemini 3.8 Flash | o3 |
|---|---|---|
| LMArena Vision | 1314 | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
| Blueprint-Bench 2 | 38.6% | — |
| Furniture Assembly | 31.7% | — |
Multilingual Gemini 3.8 Flash leads
Gemini 3.8 Flash: 58.0 (#5), o3: 51.7 (#105)
| Benchmark | Gemini 3.8 Flash | o3 |
|---|---|---|
| LMArena Non-English | 1491 | 1401 |
| LMArena Chinese | 1554 | 1437 |
| LMArena French | 1498 | 1430 |
| LMArena German | 1493 | 1420 |
| LMArena Japanese | 1502 | 1403 |
| LMArena Korean | 1459 | 1370 |
| LMArena Russian | 1515 | 1406 |
| LMArena Spanish | 1485 | 1395 |
Instruction Following Gemini 3.8 Flash leads
Gemini 3.8 Flash: 78.0 (#13), o3: 72.8 (#127)
| Benchmark | Gemini 3.8 Flash | o3 |
|---|---|---|
| LMArena Instruction Following | 1490 | 1368 |
| IFEval | — | 86.9% |
Long Context o3 leads
Gemini 3.8 Flash: 46.3 (#24), o3: 53.3 (#6)
| Benchmark | Gemini 3.8 Flash | o3 |
|---|---|---|
| LMArena Longer Query | 1508 | 1372 |
| Fiction.LiveBench | — | 88.9% |
| CL-bench | — | 17.8% |
Writing & Preference Gemini 3.8 Flash leads
Gemini 3.8 Flash: 72.2 (#15), o3: 63.5 (#64)
| Benchmark | Gemini 3.8 Flash | o3 |
|---|---|---|
| LMArena Text | 1499 | 1410 |
| LMArena Creative Writing | 1492 | 1359 |
| EQ-Bench Creative Writing | 1748 | 1676 |
| LMArena Multi-Turn | 1501 | 1405 |
| Short-Story Creative Writing | — | 83.9% |
| WildBench | — | 86.1% |
Frequently asked questions
Is Gemini 3.8 Flash better than o3?
Gemini 3.8 Flash is the stronger model overall, scoring 61.8 to 47.5 on the Noometry Index.
Which is cheaper, Gemini 3.8 Flash or o3?
Gemini 3.8 Flash is cheaper. It lists at $0.75 per million input tokens and $3.75 per million output tokens; o3 lists at $2 and $8.
Is Gemini 3.8 Flash or o3 better for coding?
Gemini 3.8 Flash scores higher on coding benchmarks: 59.2 versus 46.8 in the Noometry coding category.
Which has the bigger context window?
Gemini 3.8 Flash does, with 1.05M tokens against 200K.
How many benchmarks do Gemini 3.8 Flash and o3 share?
34 benchmarks have published results for both models. Gemini 3.8 Flash has 50 scored results on Noometry and o3 has 63.