Model comparison
Claude 3 Haiku vs Gemini 3 Flash Preview
Gemini 3 Flash Preview is the stronger model overall, scoring 52.3 to 25.9 on the Noometry Index.
Last verified . 25 shared benchmarks.
GoogleSummary
- They share 25 benchmarks with published results for both. Claude 3 Haiku scores higher in 0 categories and Gemini 3 Flash Preview in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Gemini 3 Flash Preview leads 51.7 to 9.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.8% for Claude 3 Haiku and 95.6% for Gemini 3 Flash Preview.
Side by side
| Claude 3 Haiku | Gemini 3 Flash Preview | |
|---|---|---|
| Provider | Anthropic | |
| Noometry Index | 25.9 | 52.3 |
| Released | 2024-03-07 | 2025-12-17 |
| Weights | Proprietary | Proprietary |
| Context window | — | 1.05M |
| Max output | — | 66K |
| Input $ / M tokens | — | $0.50 |
| Output $ / M tokens | — | $3 |
| Results tracked | 37 | 59 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3 Flash Preview leads
Claude 3 Haiku: 26.4 (#325), Gemini 3 Flash Preview: 50.9 (#42)
| Benchmark | Claude 3 Haiku | Gemini 3 Flash Preview |
|---|---|---|
| WeirdML | 9.8% | 61.6% |
| LMArena Coding | 1199 | 1460 |
| SWE-bench Verified | — | 75.4% |
| SWE-bench Verified (bash only) | — | 75.8% |
| LMArena WebDev | — | 1439 |
| SWE-bench Multilingual | — | 72.7% |
| GSO | — | 9.8% |
| BigCodeBench Instruct | 39.4% | — |
| BigCodeBench Complete | 50.1% | — |
| CadEval | 12% | — |
| ALE-Bench | — | 1,367 |
| HumanEval+ | 68.9% | — |
| MBPP+ | 68.8% | — |
Agentic & Tool Use Not comparable
Claude 3 Haiku: —, Gemini 3 Flash Preview: 38.7 (#29)
| Benchmark | Claude 3 Haiku | Gemini 3 Flash Preview |
|---|---|---|
| Terminal-Bench | — | 64.3% |
| τ²-bench Airline | — | 82.5% |
| τ²-bench Banking | — | 27.3% |
| τ²-bench Retail | — | 76.8% |
| τ²-bench Telecom | — | 91.2% |
| DeepResearch Bench | — | 49.8% |
| BALROG | — | 48.1% |
| GDP.pdf | — | 10% |
| LMArena Search | — | 1198 |
| Vending-Bench 2 | — | 3,635 |
Reasoning Gemini 3 Flash Preview leads
Claude 3 Haiku: 16.3 (#307), Gemini 3 Flash Preview: 49.2 (#37)
| Benchmark | Claude 3 Haiku | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Hard Prompts | 1174 | 1465 |
| DTBench | 50.1% | 89.1% |
| LMCA | 8.8% | 43.1% |
| Epoch Capabilities Index | 118.35 | 151.8 |
| ForecastBench | 53.2 | 58.5 |
| ARC-AGI-2 | — | 33.6% |
| SimpleBench | — | 61.1% |
| Kagi LLM Benchmark | 34.2% | — |
| NYT Connections (extended) | — | 83.1% |
| ARC-AGI-1 | — | 84.7% |
| Chess Puzzles | — | 40% |
| Mystery Game Puzzles | — | 26% |
| WinoGrande | 74.2% | — |
Math Gemini 3 Flash Preview leads
Claude 3 Haiku: 9.8 (#319), Gemini 3 Flash Preview: 51.7 (#55)
| Benchmark | Claude 3 Haiku | Gemini 3 Flash Preview |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.8% | 95.6% |
| LMArena Math | 1188 | 1473 |
| FrontierMath (Tiers 1-3) | — | 51.2% |
| FrontierMath Tier 4 | — | 17.1% |
| MathArena Final-Answer Competitions | — | 67.6% |
| ProofBench | — | 15% |
| MATH Level 5 | 14.9% | — |
| FrontierMath (Feb 2025 set) | — | 35.6% |
| FrontierMath Tier 4 (v1) | — | 4.2% |
Knowledge Gemini 3 Flash Preview leads
Claude 3 Haiku: 17.3 (#285), Gemini 3 Flash Preview: 58.8 (#33)
| Benchmark | Claude 3 Haiku | Gemini 3 Flash Preview |
|---|---|---|
| GPQA Diamond | 36.3% | 89.4% |
| LMArena Expert | 1148 | 1462 |
| SimpleQA Verified | — | 66.8% |
| Confabulations | 34.2% | — |
| Vectara Hallucination Rate | — | 13.5% |
| MMLU | 73.8% | — |
Multimodal Gemini 3 Flash Preview leads
Claude 3 Haiku: 23.6 (#128), Gemini 3 Flash Preview: 45.5 (#16)
| Benchmark | Claude 3 Haiku | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Vision | 950 | 1285 |
| GeoBench | — | 88% |
| VPCT | — | 72.6% |
| Blueprint-Bench 2 | — | 0% |
| LMArena Document | — | 1413 |
| ScienceQA | 72% | — |
Multilingual Gemini 3 Flash Preview leads
Claude 3 Haiku: 36.0 (#243), Gemini 3 Flash Preview: 55.7 (#27)
| Benchmark | Claude 3 Haiku | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Non-English | 1178 | 1458 |
| LMArena Chinese | 1155 | 1511 |
| LMArena French | 1195 | 1477 |
| LMArena German | 1174 | 1497 |
| LMArena Japanese | 1102 | 1489 |
| LMArena Korean | 1109 | 1443 |
| LMArena Russian | 1204 | 1480 |
| LMArena Spanish | 1166 | 1469 |
Instruction Following Gemini 3 Flash Preview leads
Claude 3 Haiku: 61.3 (#247), Gemini 3 Flash Preview: 75.7 (#56)
| Benchmark | Claude 3 Haiku | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Instruction Following | 1173 | 1437 |
Long Context Gemini 3 Flash Preview leads
Claude 3 Haiku: 36.1 (#237), Gemini 3 Flash Preview: 44.4 (#67)
| Benchmark | Claude 3 Haiku | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Longer Query | 1190 | 1452 |
Writing & Preference Gemini 3 Flash Preview leads
Claude 3 Haiku: 29.7 (#291), Gemini 3 Flash Preview: 65.5 (#45)
| Benchmark | Claude 3 Haiku | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Text | 1195 | 1466 |
| LMArena Creative Writing | 1157 | 1457 |
| LMArena Multi-Turn | 1190 | 1471 |
| EQ-Bench Creative Writing | 717 | — |
Frequently asked questions
Is Claude 3 Haiku better than Gemini 3 Flash Preview?
Gemini 3 Flash Preview is the stronger model overall, scoring 52.3 to 25.9 on the Noometry Index.
Is Claude 3 Haiku or Gemini 3 Flash Preview better for coding?
Gemini 3 Flash Preview scores higher on coding benchmarks: 50.9 versus 26.4 in the Noometry coding category.
How many benchmarks do Claude 3 Haiku and Gemini 3 Flash Preview share?
25 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and Gemini 3 Flash Preview has 59.