Codestral-22B-v0.1 vs Qwen2.5-Coder-32B-Instruct vs Vision Large
No independent capability benchmark covers every model here yet. The score uses inputs & features and context window.
Price in rating: Included · Exclude price
Mistral AI
Codestral-22B-v0.1
4/100- ECI—
- Price—
- Context33K
Alibaba (Qwen)
Qwen2.5-Coder-32B-Instruct
28/100- ECI—
- Price$0.66 / $1.00
- Context131K
Vispark
Vision Large
86/100- ECI—
- Price$7.37 / $22.11
- Context1M
Capability evidence is incomplete
There is not enough shared capability evidence to name an overall winner. The scores (Vision Large 86/100, Qwen2.5-Coder-32B-Instruct 28/100, Codestral-22B-v0.1 4/100) compare the available specifications and price. The score weighs inputs & features 60%, context window 40%. 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 priceQwen2.5-Coder-32B-InstructQwen2.5-Coder-32B-Instruct $0.745 · Vision Large $11.05 per 1M tokens (3:1 blend) · Codestral-22B-v0.1 unpriced
- Longest contextVision LargeVision Large 1,000,000 · Qwen2.5-Coder-32B-Instruct 131,072 · Codestral-22B-v0.1 32,768 tokens
- Widest inputsVision LargeCodestral-22B-v0.1: Text · Qwen2.5-Coder-32B-Instruct: Text · Vision Large: Text, Images, PDFs, Audio, Video
- Self-hostingCodestral-22B-v0.1 and Qwen2.5-Coder-32B-InstructPublishes downloadable weights (Mistral AI Non-Production License)
| Measure | Weight | Codestral-22B-v0.1 | Qwen2.5-Coder-32B-Instruct | Vision Large |
|---|---|---|---|---|
| Inputs & features | 60% | 0 | 25 | 100 |
| Context window | 40% | 11 | 32 | 64 |
| Specs & price only | 100% | 4/100 | 28/100 | 86/100 |
Left out because at least one model lacks the data: capability and price. The remaining weights were rescaled.
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | |||
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | — | — | — |
| ECI rank | — | — | — |
| Price per million tokens | |||
| Input | — | $0.66 (best) | $7.37 |
| Output | — | $1.00 (best) | $22.11 |
| Cached input | — | — | — |
| Blended (3:1) | — | $0.745 (best) | $11.05 |
| Long-context rate | — | Same rate | Same rate |
| Price source | — | Median of 7 providers | Official Vispark API |
| Limits | |||
| Context window | 32,768 tokens | 131,072 tokens | 1,000,000 tokens (best) |
| Max output | 8,192 tokens | 8,192 tokens | 65,536 tokens (best) |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | No | No | Yes |
| PDFs | No | No | Yes |
| Audio | No | No | Yes |
| Video | No | No | Yes |
| Reasoning | No | No | Yes |
| Tool calling | No | Yes | Yes |
| Structured output | No | No | Yes |
| Availability | |||
| Weights | OpenMistral AI Non-Production License | Open | Proprietary |
| API model ID | — | — | vispark/vision-large |
| API providers | — | 7 (best) | 1 |
| Released | May 29, 2024 | Nov 12, 2024 | May 15, 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.
Codestral-22B-v0.1—
Qwen2.5-Coder-32B-Instruct$8.60
Vision Large$117.92
Which should you choose?
Which is better: Codestral-22B-v0.1, Qwen2.5-Coder-32B-Instruct or Vision Large?
There is not enough shared capability evidence to name an overall winner. The scores (Vision Large 86/100, Qwen2.5-Coder-32B-Instruct 28/100, Codestral-22B-v0.1 4/100) compare the available specifications and price. The score weighs inputs & features 60%, context window 40%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.
Which is cheaper, Codestral-22B-v0.1, Qwen2.5-Coder-32B-Instruct or Vision Large?
Qwen2.5-Coder-32B-Instruct is cheaper at $0.66 input / $1.00 output per million tokens (median across 7 API providers). Vision Large costs $7.37 input / $22.11 output per million tokens (official Vispark API price). At a typical mix of three input tokens to one output token, that is $0.745 per million tokens for Qwen2.5-Coder-32B-Instruct versus $11.05 for Vision Large (15× as much). Codestral-22B-v0.1 has no published per-token price.
Which scores higher on benchmarks?
There is no independent benchmark that covers all three models yet. Codestral-22B-v0.1 has not been scored yet, Qwen2.5-Coder-32B-Instruct has not been scored yet and Vision Large has not been scored yet.
Which is better for coding?
There are no published SWE-bench Verified results for Codestral-22B-v0.1, Qwen2.5-Coder-32B-Instruct and Vision Large yet, so there is no like-for-like coding score. Test both on a sample of your own tasks. Note that Codestral-22B-v0.1 does not support tool calling, which most coding agents need.
Which has the bigger context window?
Vision Large has the largest context window at 1,000,000 tokens, against 131,072 for Qwen2.5-Coder-32B-Instruct and 32,768 for Codestral-22B-v0.1. Maximum output per response: Codestral-22B-v0.1 up to 8,192, Qwen2.5-Coder-32B-Instruct up to 8,192, Vision Large up to 65,536 tokens.
Which can read images, PDFs, audio or video?
Codestral-22B-v0.1 accepts text; Qwen2.5-Coder-32B-Instruct accepts text; Vision Large accepts text, images, PDFs, audio and video. Vision Large handles the widest range of inputs.
Are any of these open source?
Codestral-22B-v0.1 and Qwen2.5-Coder-32B-Instruct publishes its weights (Mistral AI Non-Production License) and can be self-hosted; Vision Large is proprietary.
Which is newer?
Qwen2.5-Coder-32B-Instruct is the newest, released Nov 12, 2024. Codestral-22B-v0.1 came out May 29, 2024; Vision Large came out May 15, 2024.
How do you decide the winner?
This rating uses inputs & features (60%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (40%, 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 18, 2026.