Gemma 4 E4B IT vs Ministral 3 3B vs Step 3.5 Flash 2603
Too close to call on our weighted score (Gemma 4 E4B IT 76, Ministral 3 3B 74, Step 3.5 Flash 2603 62). The right pick depends on what you value most.
GoogleGemma 4 E4B IT
76/100- ECI—
- Price$0.11 / $0.15
- Context131K
Mistral AIMinistral 3 3B
74/100- ECI—
- Price$0.10 / $0.10
- Context262K
StepFunStep 3.5 Flash 2603
62/100- ECI—
- Price$0.10 / $0.30
- Context256K
Too close to call
It is close. Our weighted score puts them within 2 points (Gemma 4 E4B IT 76/100, Ministral 3 3B 74/100, Step 3.5 Flash 2603 62/100), so choose by what matters most for your work: Ministral 3 3B on price and Ministral 3 3B for long inputs. The score weighs price 50%, inputs & features 30%, context window 20%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.
- CapabilityNot enough dataNo independent benchmark covers every model here yet
- Lowest priceMinistral 3 3BMinistral 3 3B $0.10 · Gemma 4 E4B IT $0.12 · Step 3.5 Flash 2603 $0.15 per 1M tokens (3:1 blend)
- Longest contextMinistral 3 3BMinistral 3 3B 262,144 · Step 3.5 Flash 2603 256,000 · Gemma 4 E4B IT 131,072 tokens
- Widest inputsGemma 4 E4B ITGemma 4 E4B IT: Text, Images, Audio · Ministral 3 3B: Text, Images · Step 3.5 Flash 2603: Text
- Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
| Measure | Weight | Gemma 4 E4B IT | Ministral 3 3B | Step 3.5 Flash 2603 |
|---|---|---|---|---|
| Price | 50% | 94 | 97 | 89 |
| Inputs & features | 30% | 80 | 60 | 35 |
| Context window | 20% | 24 | 37 | 36 |
| Overall | 100% | 76/100 | 74/100 | 62/100 |
Left out because at least one model lacks the data: capability. The remaining weights were rescaled.
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | Gemma 4 E4B ITGoogle | Ministral 3 3BMistral AI | Step 3.5 Flash 2603StepFun |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | — | — | — |
| ECI rank | — | — | — |
| Price per million tokens | |||
| Input | $0.11 | $0.10 (best) | $0.10 (best) |
| Output | $0.15 | $0.10 (best) | $0.30 |
| Cached input | — | — | $0.02 |
| Blended (3:1) | $0.12 | $0.10 (best) | $0.15 |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Median of 2 providers | Median of 1 providers | Official StepFun (Global) API |
| Limits | |||
| Context window | 131,072 tokens | 262,144 tokens (best) | 256,000 tokens |
| Max output | 8,192 tokens | 262,144 tokens (best) | 256,000 tokens |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | No |
| PDFs | No | No | No |
| Audio | Yes | No | No |
| Video | No | No | No |
| Reasoning | Yes | No | Yeslow · high |
| Tool calling | Yes | Yes | Yes |
| Structured output | Yes | Yes | No |
| Availability | |||
| Weights | Open | OpenApache 2.0 | Open |
| API model ID | — | — | step-3.5-flash-2603 |
| API providers | 2 | 1 | 3 (best) |
| Released | Apr 2, 2026 | Dec 2, 2025 | Apr 2, 2026 |
| Knowledge cutoff | — | — | Jan 2025 |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
- Gemma 4 E4B IT$1.40
- Ministral 3 3B$1.20
- Step 3.5 Flash 2603$1.60
Which should you choose?
Which is better: Gemma 4 E4B IT, Ministral 3 3B or Step 3.5 Flash 2603?
It is close. Our weighted score puts them within 2 points (Gemma 4 E4B IT 76/100, Ministral 3 3B 74/100, Step 3.5 Flash 2603 62/100), so choose by what matters most for your work: Ministral 3 3B on price and Ministral 3 3B for long inputs. The score weighs price 50%, inputs & features 30%, context window 20%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.
Which is cheaper, Gemma 4 E4B IT, Ministral 3 3B or Step 3.5 Flash 2603?
Ministral 3 3B is cheaper at $0.10 input / $0.10 output per million tokens (median across 1 API provider). Gemma 4 E4B IT costs $0.11 input / $0.15 output per million tokens (median across 2 API providers); Step 3.5 Flash 2603 costs $0.10 input / $0.30 output per million tokens (official StepFun (Global) API price). At a typical mix of three input tokens to one output token, that is $0.10 per million tokens for Ministral 3 3B versus $0.12 for Gemma 4 E4B IT (1.2× as much) and $0.15 for Step 3.5 Flash 2603 (1.5× as much).
Which scores higher on benchmarks?
There is no independent benchmark that covers all three models yet. Gemma 4 E4B IT has not been scored yet, Ministral 3 3B has not been scored yet and Step 3.5 Flash 2603 has not been scored yet.
Which is better for coding?
There are no published SWE-bench Verified results for Gemma 4 E4B IT, Ministral 3 3B and Step 3.5 Flash 2603 yet, so there is no like-for-like coding score. Test both on a sample of your own tasks. All three support tool calling for agent workflows.
Which has the bigger context window?
Ministral 3 3B has the largest context window at 262,144 tokens, against 256,000 for Step 3.5 Flash 2603 and 131,072 for Gemma 4 E4B IT. Maximum output per response: Gemma 4 E4B IT up to 8,192, Ministral 3 3B up to 262,144, Step 3.5 Flash 2603 up to 256,000 tokens.
Which can read images, PDFs, audio or video?
Gemma 4 E4B IT accepts text, images and audio; Ministral 3 3B accepts text and images; Step 3.5 Flash 2603 accepts text. Gemma 4 E4B IT handles the widest range of inputs.
Are any of these open source?
Yes, all three publish their weights (Apache 2.0), so you can self-host them.
Which is newer?
Gemma 4 E4B IT is the newest, released Apr 2, 2026. Step 3.5 Flash 2603 came out Apr 2, 2026; Ministral 3 3B came out Dec 2, 2025. Knowledge cutoff: Step 3.5 Flash 2603 Jan 2025.
How do you decide the winner?
Each model gets a 0–100 score on capability (50%, independent benchmark results); price (25%, blended price per million tokens (3 input : 1 output), log scale); inputs & features (15%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (10%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Oct 4, 2026.
