LongCat-2.5-Preview vs Gemma 4 12B IT vs Step 5 Preview
Gemma 4 12B IT comes out ahead, 75 to 57 and 52 on our weighted score, and it is the cheaper option too.
Meituan (LongCat)LongCat-2.5-Preview
52/100- ECI—
- Price$0.75 / $2.95
- Context1M
- Our pick
GoogleGemma 4 12B IT
75/100- ECI—
- Price$0.075 / $0.275
- Context262K
StepFunStep 5 Preview
57/100- ECI—
- Price$1.00 / $2.70
- Context1M
Gemma 4 12B IT is our pick
Gemma 4 12B IT is the better all-round choice, scoring 75/100 against Step 5 Preview (57) and LongCat-2.5-Preview (52). It leads on price. Step 5 Preview wins on inputs & features. 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 priceGemma 4 12B ITGemma 4 12B IT $0.125 · LongCat-2.5-Preview $1.30 · Step 5 Preview $1.43 per 1M tokens (3:1 blend)
- Longest contextLongCat-2.5-Preview and Step 5 PreviewLongCat-2.5-Preview 1,000,000 · Step 5 Preview 1,000,000 · Gemma 4 12B IT 262,144 tokens
- Widest inputsStep 5 PreviewLongCat-2.5-Preview: Text, Images · Gemma 4 12B IT: Text, Images · Step 5 Preview: Text, Images, Video
- Self-hostingGemma 4 12B ITPublishes downloadable weights
| Measure | Weight | LongCat-2.5-Preview | Gemma 4 12B IT | Step 5 Preview |
|---|---|---|---|---|
| Price | 50% | 44 | 93 | 43 |
| Inputs & features | 30% | 60 | 70 | 80 |
| Context window | 20% | 60 | 37 | 60 |
| Overall | 100% | 52/100 | 75/100 | 57/100 |
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 | LongCat-2.5-PreviewMeituan (LongCat) | Gemma 4 12B ITGoogle | Step 5 PreviewStepFun |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | — | — | — |
| ECI rank | — | — | — |
| Price per million tokens | |||
| Input | $0.75 | $0.075 (best) | $1.00 |
| Output | $2.95 | $0.275 (best) | $2.70 |
| Cached input | — | — | $0.05 |
| Blended (3:1) | $1.30 | $0.125 (best) | $1.43 |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Median of 1 providers | Median of 2 providers | Official StepFun (Global) API |
| Limits | |||
| Context window | 1,000,000 tokens (best) | 262,144 tokens | 1,000,000 tokens (best) |
| Max output | 131,072 tokens | 32,768 tokens | 1,000,000 tokens (best) |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | Yes |
| PDFs | No | No | No |
| Audio | No | No | No |
| Video | No | No | Yes |
| Reasoning | Yes | Yes | Yeslow · medium · high |
| Tool calling | Yes | Yes | Yes |
| Structured output | No | Yes | Yes |
| Availability | |||
| Weights | Proprietary | Open | Proprietary |
| API model ID | — | — | step-5-preview |
| API providers | 2 | 2 | 7 (best) |
| Released | Sep 25, 2026 | Jun 9, 2026 | Sep 16, 2026 |
| Knowledge cutoff | — | — | — |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
- LongCat-2.5-Preview$13.40
- Gemma 4 12B IT$1.30
- Step 5 Preview$15.40
Which should you choose?
Which is better: LongCat-2.5-Preview, Gemma 4 12B IT or Step 5 Preview?
Gemma 4 12B IT is the better all-round choice, scoring 75/100 against Step 5 Preview (57) and LongCat-2.5-Preview (52). It leads on price. Step 5 Preview wins on inputs & features. 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, LongCat-2.5-Preview, Gemma 4 12B IT or Step 5 Preview?
Gemma 4 12B IT is cheaper at $0.075 input / $0.275 output per million tokens (median across 2 API providers). LongCat-2.5-Preview costs $0.75 input / $2.95 output per million tokens (median across 1 API provider); Step 5 Preview costs $1.00 input / $2.70 output per million tokens (official StepFun (Global) API price). At a typical mix of three input tokens to one output token, that is $0.125 per million tokens for Gemma 4 12B IT versus $1.30 for LongCat-2.5-Preview (10× as much) and $1.43 for Step 5 Preview (11× as much).
Which scores higher on benchmarks?
There is no independent benchmark that covers all three models yet. LongCat-2.5-Preview has not been scored yet, Gemma 4 12B IT has not been scored yet and Step 5 Preview has not been scored yet.
Which is better for coding?
There are no published SWE-bench Verified results for LongCat-2.5-Preview, Gemma 4 12B IT and Step 5 Preview 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?
LongCat-2.5-Preview and Step 5 Preview have the largest context windows (1,000,000 and 1,000,000 tokens), against 262,144 for Gemma 4 12B IT. Maximum output per response: LongCat-2.5-Preview up to 131,072, Gemma 4 12B IT up to 32,768, Step 5 Preview up to 1,000,000 tokens.
Which can read images, PDFs, audio or video?
LongCat-2.5-Preview accepts text and images; Gemma 4 12B IT accepts text and images; Step 5 Preview accepts text, images and video. Step 5 Preview handles the widest range of inputs.
Are any of these open source?
Gemma 4 12B IT publishes its weights and can be self-hosted; LongCat-2.5-Preview and Step 5 Preview is proprietary.
Which is newer?
LongCat-2.5-Preview is the newest, released Sep 25, 2026. Step 5 Preview came out Sep 16, 2026; Gemma 4 12B IT came out Jun 9, 2026.
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.
