Aya Expanse 32B vs Voxtral Small 24B 2507 vs Qwen2.5-Coder-32B-Instruct
Voxtral Small 24B 2507 comes out ahead, 55 to 39 and 32 on our weighted score, and it is the cheaper option too.
CohereAya Expanse 32B
32/100- ECI—
- Price$0.50 / $1.50
- Context128K
- Our pick
Mistral AIVoxtral Small 24B 2507
55/100- ECI—
- Price$0.10 / $0.40
- Context33K
Alibaba (Qwen)Qwen2.5-Coder-32B-Instruct
39/100- ECI—
- Price$0.66 / $1.00
- Context131K
Voxtral Small 24B 2507 is our pick
Voxtral Small 24B 2507 is the better all-round choice, scoring 55/100 against Qwen2.5-Coder-32B-Instruct (39) and Aya Expanse 32B (32). 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 priceVoxtral Small 24B 2507Voxtral Small 24B 2507 $0.175 · Qwen2.5-Coder-32B-Instruct $0.745 · Aya Expanse 32B $0.75 per 1M tokens (3:1 blend)
- Longest contextQwen2.5-Coder-32B-InstructQwen2.5-Coder-32B-Instruct 131,072 · Aya Expanse 32B 128,000 · Voxtral Small 24B 2507 32,768 tokens
- Widest inputsVoxtral Small 24B 2507Aya Expanse 32B: Text · Voxtral Small 24B 2507: Text, Audio · 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 | Voxtral Small 24B 2507 | Qwen2.5-Coder-32B-Instruct |
|---|---|---|---|---|
| Price | 50% | 53 | 82 | 54 |
| Inputs & features | 30% | 0 | 45 | 25 |
| Context window | 20% | 24 | 0 | 25 |
| Overall | 100% | 32/100 | 55/100 | 39/100 |
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 | Voxtral Small 24B 2507Mistral AI | Qwen2.5-Coder-32B-InstructAlibaba (Qwen) |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | — | — | — |
| ECI rank | — | — | — |
| Price per million tokens | |||
| Input | $0.50 | $0.10 (best) | $0.66 |
| Output | $1.50 | $0.40 (best) | $1.00 |
| Cached input | — | — | — |
| Blended (3:1) | $0.75 | $0.175 (best) | $0.745 |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Official Cohere API | Official Mistral API | Median of 7 providers |
| Limits | |||
| Context window | 128,000 tokens | 32,768 tokens | 131,072 tokens (best) |
| Max output | 4,000 tokens | 32,768 tokens (best) | 8,192 tokens |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | No | No | No |
| PDFs | No | No | No |
| Audio | No | Yes | No |
| Video | No | No | No |
| Reasoning | No | No | No |
| Tool calling | No | Yes | Yes |
| Structured output | No | Yes | No |
| Availability | |||
| Weights | OpenCC-BY-NC-4.0 | OpenApache 2.0 | Open |
| API model ID | c4ai-aya-expanse-32b | voxtral-small-latest | — |
| API providers | 2 | 10 (best) | 7 |
| Released | Oct 24, 2024 | Jul 15, 2025 | Nov 12, 2024 |
| 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.
- Aya Expanse 32B$8.00
- Voxtral Small 24B 2507$1.80
- Qwen2.5-Coder-32B-Instruct$8.60
Which should you choose?
Which is better: Aya Expanse 32B, Voxtral Small 24B 2507 or Qwen2.5-Coder-32B-Instruct?
Voxtral Small 24B 2507 is the better all-round choice, scoring 55/100 against Qwen2.5-Coder-32B-Instruct (39) and Aya Expanse 32B (32). 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, Voxtral Small 24B 2507 or Qwen2.5-Coder-32B-Instruct?
Voxtral Small 24B 2507 is cheaper at $0.10 input / $0.40 output per million tokens (official Mistral API price). Qwen2.5-Coder-32B-Instruct costs $0.66 input / $1.00 output per million tokens (median across 7 API providers); Aya Expanse 32B costs $0.50 input / $1.50 output per million tokens (official Cohere API price). At a typical mix of three input tokens to one output token, that is $0.175 per million tokens for Voxtral Small 24B 2507 versus $0.745 for Qwen2.5-Coder-32B-Instruct (4.3× as much) and $0.75 for Aya Expanse 32B (4.3× 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, Voxtral Small 24B 2507 has not been scored yet 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, Voxtral Small 24B 2507 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?
Qwen2.5-Coder-32B-Instruct has the largest context window at 131,072 tokens, against 128,000 for Aya Expanse 32B and 32,768 for Voxtral Small 24B 2507. Maximum output per response: Aya Expanse 32B up to 4,000, Voxtral Small 24B 2507 up to 32,768, Qwen2.5-Coder-32B-Instruct up to 8,192 tokens.
Which can read images, PDFs, audio or video?
Aya Expanse 32B accepts text; Voxtral Small 24B 2507 accepts text and audio; Qwen2.5-Coder-32B-Instruct accepts text. Voxtral Small 24B 2507 handles the widest range of inputs.
Are any of these open source?
Yes, all three publish their weights (CC-BY-NC-4.0 and Apache 2.0), so you can self-host them.
Which is newer?
Voxtral Small 24B 2507 is the newest, released Jul 15, 2025. Qwen2.5-Coder-32B-Instruct came out Nov 12, 2024; Aya Expanse 32B came out Oct 24, 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 Nov 2, 2026.
