Aya Expanse 32B vs Gemma 3 12B IT vs Qwen2.5-Coder-32B-Instruct
Gemma 3 12B IT comes out ahead, 70 to 45 and 33 on our weighted score, and it is the cheaper option too.
CohereAya Expanse 32B
33/100- ECI—
- Price$0.50 / $1.50
- Context128K
- Our pick
GoogleGemma 3 12B IT
70/100- ECI123.5
- Price$0.05 / $0.15
- Context131K
Alibaba (Qwen)Qwen2.5-Coder-32B-Instruct
45/100- ECI—
- Price$0.43 / $0.60
- Context131K
Gemma 3 12B IT is our pick
Gemma 3 12B IT is the better all-round choice, scoring 70/100 against Qwen2.5-Coder-32B-Instruct (45) and Aya Expanse 32B (33). It leads on price and 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 3 12B ITGemma 3 12B IT $0.075 · Qwen2.5-Coder-32B-Instruct $0.473 · Aya Expanse 32B $0.75 per 1M tokens (3:1 blend)
- Longest contextGemma 3 12B IT and Qwen2.5-Coder-32B-InstructGemma 3 12B IT 131,072 · Qwen2.5-Coder-32B-Instruct 131,072 · Aya Expanse 32B 128,000 tokens
- Widest inputsGemma 3 12B ITAya Expanse 32B: Text · Gemma 3 12B IT: Text, Images · Qwen2.5-Coder-32B-Instruct: Text
- Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
| Measure | Weight | Aya Expanse 32B | Gemma 3 12B IT | Qwen2.5-Coder-32B-Instruct |
|---|---|---|---|---|
| Price | 50% | 56 | 100 | 65 |
| Inputs & features | 30% | 0 | 50 | 25 |
| Context window | 20% | 24 | 24 | 24 |
| Overall | 100% | 33/100 | 70/100 | 45/100 |
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 | Aya Expanse 32BCohere | Gemma 3 12B ITGoogle | Qwen2.5-Coder-32B-InstructAlibaba (Qwen) |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | — | 123.5 | — |
| ECI rank | — | #138 of 148 | — |
| GPQA DiamondGraduate-level science questions | — | 39.5% | — |
| OTIS Mock AIME 2024–2025Competition mathematics | — | 16.7% | — |
| Price per million tokens | |||
| Input | $0.50 | $0.05 (best) | $0.43 |
| Output | $1.50 | $0.15 (best) | $0.60 |
| Cached input | — | — | — |
| Blended (3:1) | $0.75 | $0.075 (best) | $0.473 |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Median of 1 providers | Median of 7 providers | Median of 4 providers |
| Limits | |||
| Context window | 128,000 tokens | 131,072 tokens (best) | 131,072 tokens (best) |
| Max output | 4,000 tokens | 131,072 tokens (best) | 8,192 tokens |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | No | Yes | No |
| PDFs | No | No | No |
| Audio | No | No | No |
| Video | No | No | No |
| Reasoning | No | No | No |
| Tool calling | No | Yes | Yes |
| Structured output | No | No | No |
| Availability | |||
| Weights | OpenCC-BY-NC-4.0 | Open | Open |
| API model ID | c4ai-aya-expanse-32b | — | — |
| API providers | 2 | 7 (best) | 4 |
| Released | Oct 24, 2024 | Mar 12, 2025 | Nov 12, 2024 |
| Knowledge cutoff | — | Aug 2024 | — |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
- Aya Expanse 32B$8.00
- Gemma 3 12B IT$0.80
- Qwen2.5-Coder-32B-Instruct$5.50
Which should you choose?
Which is better: Aya Expanse 32B, Gemma 3 12B IT or Qwen2.5-Coder-32B-Instruct?
Gemma 3 12B IT is the better all-round choice, scoring 70/100 against Qwen2.5-Coder-32B-Instruct (45) and Aya Expanse 32B (33). It leads on price and 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, Aya Expanse 32B, Gemma 3 12B IT or Qwen2.5-Coder-32B-Instruct?
Gemma 3 12B IT is cheaper at $0.05 input / $0.15 output per million tokens (median across 7 API providers). Qwen2.5-Coder-32B-Instruct costs $0.43 input / $0.60 output per million tokens (median across 4 API providers); Aya Expanse 32B costs $0.50 input / $1.50 output per million tokens (median across 1 API provider). At a typical mix of three input tokens to one output token, that is $0.075 per million tokens for Gemma 3 12B IT versus $0.473 for Qwen2.5-Coder-32B-Instruct (6.3× as much) and $0.75 for Aya Expanse 32B (10× as much).
Which scores higher on benchmarks?
There is no independent benchmark that covers all three models yet. Aya Expanse 32B has not been scored yet, Gemma 3 12B IT has an ECI of 123.5 and Qwen2.5-Coder-32B-Instruct has not been scored yet.
Which is better for coding?
There are no published SWE-bench Verified results for Aya Expanse 32B, Gemma 3 12B IT and Qwen2.5-Coder-32B-Instruct yet, so there is no like-for-like coding score. Test both on a sample of your own tasks. Note that Aya Expanse 32B does not support tool calling, which most coding agents need.
Which has the bigger context window?
Gemma 3 12B IT and Qwen2.5-Coder-32B-Instruct have the largest context windows (131,072 and 131,072 tokens), against 128,000 for Aya Expanse 32B. Maximum output per response: Aya Expanse 32B up to 4,000, Gemma 3 12B IT up to 131,072, Qwen2.5-Coder-32B-Instruct up to 8,192 tokens.
Which can read images, PDFs, audio or video?
Aya Expanse 32B accepts text; Gemma 3 12B IT accepts text and images; Qwen2.5-Coder-32B-Instruct accepts text. Gemma 3 12B IT handles the widest range of inputs.
Are any of these open source?
Yes, all three publish their weights (CC-BY-NC-4.0), so you can self-host them.
Which is newer?
Gemma 3 12B IT is the newest, released Mar 12, 2025. Qwen2.5-Coder-32B-Instruct came out Nov 12, 2024; Aya Expanse 32B came out Oct 24, 2024. Knowledge cutoff: Gemma 3 12B IT Aug 2024.
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.
