Model comparison
Gemini 3.8 Flash vs Step 5 Preview
Gemini 3.8 Flash is the stronger model overall, scoring 61.8 to 47.9 on the Noometry Index.
Last verified . 18 shared benchmarks.
Google
StepFunSummary
- They share 18 benchmarks with published results for both. Gemini 3.8 Flash scores higher in 8 categories and Step 5 Preview in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3.8 Flash leads 76.9 to 40.0.
- The biggest single-benchmark swing is ProofBench: 48% for Gemini 3.8 Flash and 42% for Step 5 Preview.
- Both cost about the same: $0.75 input and $3.75 output per million tokens.
- Gemini 3.8 Flash accepts more context: 1.05M tokens versus 1.02M.
Side by side
| Gemini 3.8 Flash | Step 5 Preview | |
|---|---|---|
| Provider | StepFun | |
| Noometry Index | 61.8 | 47.9 |
| Released | 2026-09-02 | 2026-09-16 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 1.02M |
| Max output | 66K | 66K |
| Input $ / M tokens | $0.75 | $1 |
| Output $ / M tokens | $3.75 | $2.70 |
| Results tracked | 50 | 18 |
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), Step 5 Preview: 51.8 (#36)
| Benchmark | Gemini 3.8 Flash | Step 5 Preview |
|---|---|---|
| LMArena WebDev | 1584 | 1564 |
| SciCode | 56.6% | 58.9% |
| LMArena Coding | 1510 | 1480 |
| DeepSWE | 73.8% | — |
| FrontierCode | 41.2% | — |
| CursorBench | 39.6% | — |
| FrontierSWE | 19.6% | — |
| WeirdML | 84.8% | — |
| ALE-Bench | 1,270 | — |
Agentic & Tool Use Not comparable
Gemini 3.8 Flash: 41.8 (#21), Step 5 Preview: —
| Benchmark | Gemini 3.8 Flash | Step 5 Preview |
|---|---|---|
| APEX-Agents | 64.3% | — |
| Remote Labor Index | 5.8% | — |
| GDP.pdf | 23.4% | — |
| Vending-Bench 2 | 5,094 | — |
Reasoning Gemini 3.8 Flash leads
Gemini 3.8 Flash: 76.9 (#5), Step 5 Preview: 40.0 (#57)
| Benchmark | Gemini 3.8 Flash | Step 5 Preview |
|---|---|---|
| CritPt | 18.3% | 20.9% |
| LMArena Hard Prompts | 1508 | 1465 |
| ARC-AGI-2 | 89.2% | — |
| NYT Connections (extended) | 97.4% | — |
| ARC-AGI-1 | 98.5% | — |
| Chess Puzzles | 61% | — |
| Mystery Game Puzzles | 47% | — |
| DTBench | 95.7% | — |
| LMCA | 52.9% | — |
| Surface Evolver Bench | 76.9% | — |
| Epoch Capabilities Index | 156.71 | — |
Math Gemini 3.8 Flash leads
Gemini 3.8 Flash: 65.3 (#28), Step 5 Preview: 46.1 (#72)
| Benchmark | Gemini 3.8 Flash | Step 5 Preview |
|---|---|---|
| ProofBench | 48% | 42% |
| LMArena Math | 1528 | 1470 |
| FrontierMath (Tiers 1-3) | 68.4% | — |
| FrontierMath Tier 4 | 22% | — |
| OTIS Mock AIME 2024-2025 | 98.9% | — |
Knowledge Gemini 3.8 Flash leads
Gemini 3.8 Flash: 74.8 (#2), Step 5 Preview: 41.2 (#112)
| Benchmark | Gemini 3.8 Flash | Step 5 Preview |
|---|---|---|
| LMArena Expert | 1524 | 1470 |
| GPQA Diamond | 95.4% | — |
| Humanity's Last Exam | 44.5% | — |
| SimpleQA Verified | 69.7% | — |
Multimodal Too close to call
Gemini 3.8 Flash: 40.7 (#45), Step 5 Preview: 41.0 (#41)
| Benchmark | Gemini 3.8 Flash | Step 5 Preview |
|---|---|---|
| LMArena Vision | 1314 | 1267 |
| Blueprint-Bench 2 | 38.6% | — |
| Furniture Assembly | 31.7% | — |
Multilingual Gemini 3.8 Flash leads
Gemini 3.8 Flash: 58.0 (#5), Step 5 Preview: 53.8 (#54)
| Benchmark | Gemini 3.8 Flash | Step 5 Preview |
|---|---|---|
| LMArena Non-English | 1491 | 1432 |
| LMArena Chinese | 1554 | 1519 |
| LMArena Russian | 1515 | 1436 |
| LMArena Spanish | 1485 | 1439 |
| LMArena French | 1498 | — |
| LMArena German | 1493 | — |
| LMArena Japanese | 1502 | — |
| LMArena Korean | 1459 | — |
Instruction Following Gemini 3.8 Flash leads
Gemini 3.8 Flash: 78.0 (#13), Step 5 Preview: 76.0 (#50)
| Benchmark | Gemini 3.8 Flash | Step 5 Preview |
|---|---|---|
| LMArena Instruction Following | 1490 | 1444 |
Long Context Gemini 3.8 Flash leads
Gemini 3.8 Flash: 46.3 (#24), Step 5 Preview: 44.5 (#64)
| Benchmark | Gemini 3.8 Flash | Step 5 Preview |
|---|---|---|
| LMArena Longer Query | 1508 | 1455 |
Writing & Preference Gemini 3.8 Flash leads
Gemini 3.8 Flash: 72.2 (#15), Step 5 Preview: 62.8 (#71)
| Benchmark | Gemini 3.8 Flash | Step 5 Preview |
|---|---|---|
| LMArena Text | 1499 | 1442 |
| LMArena Creative Writing | 1492 | 1410 |
| LMArena Multi-Turn | 1501 | 1447 |
| EQ-Bench Creative Writing | 1748 | — |
Frequently asked questions
Is Gemini 3.8 Flash better than Step 5 Preview?
Gemini 3.8 Flash is the stronger model overall, scoring 61.8 to 47.9 on the Noometry Index.
Which is cheaper, Gemini 3.8 Flash or Step 5 Preview?
Step 5 Preview is cheaper. It lists at $1 per million input tokens and $2.70 per million output tokens; Gemini 3.8 Flash lists at $0.75 and $3.75.
Is Gemini 3.8 Flash or Step 5 Preview better for coding?
Gemini 3.8 Flash scores higher on coding benchmarks: 59.2 versus 51.8 in the Noometry coding category.
Which has the bigger context window?
Gemini 3.8 Flash does, with 1.05M tokens against 1.02M.
How many benchmarks do Gemini 3.8 Flash and Step 5 Preview share?
18 benchmarks have published results for both models. Gemini 3.8 Flash has 50 scored results on Noometry and Step 5 Preview has 18.