Model comparison
Gemini 2.5 Flash vs o3
o3 is the stronger model overall, scoring 47.5 to 39.3 on the Noometry Index. Gemini 2.5 Flash costs 4.1× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Last verified . 48 shared benchmarks.
GoogleSummary
- They share 48 benchmarks with published results for both. Gemini 2.5 Flash scores higher in 3 categories and o3 in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where o3 leads 54.6 to 36.4.
- The biggest single-benchmark swing is GPQA (HELM): 39% for Gemini 2.5 Flash and 75.3% for o3.
- Gemini 2.5 Flash is cheaper at $0.30 / $2.50 per million input/output tokens, against $2 / $8 for o3.
- Gemini 2.5 Flash accepts more context: 1.05M tokens versus 200K.
Side by side
| Gemini 2.5 Flash | o3 | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 39.3 | 47.5 |
| Released | 2025-04-17 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 200K |
| Max output | 66K | 100K |
| Input $ / M tokens | $0.30 | $2 |
| Output $ / M tokens | $2.50 | $8 |
| Results tracked | 54 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3 leads
Gemini 2.5 Flash: 35.8 (#220), o3: 46.8 (#64)
| Benchmark | Gemini 2.5 Flash | o3 |
|---|---|---|
| SWE-bench Verified (bash only) | 28.7% | 58.4% |
| Aider Polyglot | 55.1% | 81.3% |
| WeirdML | 41.9% | 52.4% |
| LMArena Coding | 1424 | 1408 |
| ALE-Bench | 661.88 | 933.55 |
| SWE-bench Verified | — | 62.3% |
| GSO | — | 8.8% |
| CadEval | — | 74% |
Agentic & Tool Use o3 leads
Gemini 2.5 Flash: 30.8 (#74), o3: 34.5 (#44)
| Benchmark | Gemini 2.5 Flash | o3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 56.2% | 63% |
| Terminal-Bench | 17.1% | — |
| GDPval | — | 30.8% |
| TheAgentCompany | 41.1% | — |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| BALROG | 33.5% | — |
| LMArena Search | — | 1144 |
| METR Time Horizons | — | 65.4% |
| Vending-Bench 2 | 548.84 | — |
Reasoning o3 leads
Gemini 2.5 Flash: 18.1 (#286), o3: 32.0 (#78)
| Benchmark | Gemini 2.5 Flash | o3 |
|---|---|---|
| ARC-AGI-2 | 2.5% | 6.5% |
| SimpleBench | 41.2% | 53.1% |
| Kagi LLM Benchmark | 56.8% | 67.6% |
| ARC-AGI-1 | 33.3% | 60.8% |
| CritPt | 1.1% | 1.4% |
| EnigmaEval | 2.7% | 13.1% |
| LMArena Hard Prompts | 1422 | 1402 |
| DTBench | 76.5% | 84.8% |
| LMCA | 27.5% | 39.7% |
| Epoch Capabilities Index | 143.03 | 146.86 |
| ForecastBench | 60.6 | 62.5 |
| Chess Puzzles | — | 38% |
| Mystery Game Puzzles | — | 29% |
Math o3 leads
Gemini 2.5 Flash: 39.9 (#98), o3: 50.2 (#58)
| Benchmark | Gemini 2.5 Flash | o3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 73.1% | 84.4% |
| Omni-MATH | 38.5% | 71.4% |
| LMArena Math | 1415 | 1426 |
| FrontierMath (Feb 2025 set) | 4.8% | 18.7% |
| FrontierMath Tier 4 (v1) | 4.2% | 2.1% |
| FrontierMath (Tiers 1-3) | — | 33.3% |
| MATH Level 5 | — | 97.8% |
Knowledge o3 leads
Gemini 2.5 Flash: 36.4 (#168), o3: 54.6 (#52)
| Benchmark | Gemini 2.5 Flash | o3 |
|---|---|---|
| Humanity's Last Exam | 12.1% | 20.3% |
| MMLU-Pro | 63.9% | 85.9% |
| Confabulations | 16.8% | 14.4% |
| GPQA (HELM) | 39% | 75.3% |
| LMArena Expert | 1426 | 1402 |
| GPQA Diamond | — | 81.8% |
| SimpleQA Verified | — | 49.4% |
| Vectara Hallucination Rate | 7.8% | — |
Multimodal Too close to call
Gemini 2.5 Flash: 41.8 (#32), o3: 41.4 (#36)
| Benchmark | Gemini 2.5 Flash | o3 |
|---|---|---|
| LMArena Vision | 1253 | 1214 |
| GeoBench | 76% | 74% |
| VPCT | 46.2% | 52% |
| SpatialViz-Bench | 36.9% | — |
Multilingual Too close to call
Gemini 2.5 Flash: 52.3 (#88), o3: 51.7 (#105)
| Benchmark | Gemini 2.5 Flash | o3 |
|---|---|---|
| LMArena Non-English | 1409 | 1401 |
| LMArena Chinese | 1450 | 1437 |
| LMArena French | 1433 | 1430 |
| LMArena German | 1418 | 1420 |
| LMArena Japanese | 1405 | 1403 |
| LMArena Korean | 1385 | 1370 |
| LMArena Russian | 1415 | 1406 |
| LMArena Spanish | 1421 | 1395 |
Instruction Following Gemini 2.5 Flash leads
Gemini 2.5 Flash: 75.7 (#54), o3: 72.8 (#127)
| Benchmark | Gemini 2.5 Flash | o3 |
|---|---|---|
| IFEval | 89.8% | 86.9% |
| LMArena Instruction Following | 1405 | 1368 |
Long Context o3 leads
Gemini 2.5 Flash: 47.5 (#17), o3: 53.3 (#6)
| Benchmark | Gemini 2.5 Flash | o3 |
|---|---|---|
| Fiction.LiveBench | 77.8% | 88.9% |
| LMArena Longer Query | 1419 | 1372 |
| CL-bench | — | 17.8% |
Writing & Preference o3 leads
Gemini 2.5 Flash: 53.8 (#157), o3: 63.5 (#64)
| Benchmark | Gemini 2.5 Flash | o3 |
|---|---|---|
| LMArena Text | 1417 | 1410 |
| LMArena Creative Writing | 1400 | 1359 |
| Short-Story Creative Writing | 76.5% | 83.9% |
| EQ-Bench Creative Writing | 1137 | 1676 |
| WildBench | 81.7% | 86.1% |
| LMArena Multi-Turn | 1408 | 1405 |
Frequently asked questions
Is Gemini 2.5 Flash better than o3?
o3 is the stronger model overall, scoring 47.5 to 39.3 on the Noometry Index. Gemini 2.5 Flash costs 4.1× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Which is cheaper, Gemini 2.5 Flash or o3?
Gemini 2.5 Flash is cheaper. It lists at $0.30 per million input tokens and $2.50 per million output tokens; o3 lists at $2 and $8.
Is Gemini 2.5 Flash or o3 better for coding?
o3 scores higher on coding benchmarks: 46.8 versus 35.8 in the Noometry coding category.
Which has the bigger context window?
Gemini 2.5 Flash does, with 1.05M tokens against 200K.
How many benchmarks do Gemini 2.5 Flash and o3 share?
48 benchmarks have published results for both models. Gemini 2.5 Flash has 54 scored results on Noometry and o3 has 63.