Vision Medium vs Qwen2.5-VL 7B Instruct vs Codestral
Vision Medium comes out ahead, 84 to 40 and 29 on our weighted score.
- Our pick
VisparkVision Medium
84/100- ECI—
- Price—
- Context1M
Alibaba (Qwen)Qwen2.5-VL 7B Instruct
40/100- ECI—
- Price$0.35 / $1.05
- Context131K
Mistral AICodestral
29/100- ECI—
- Price$0.30 / $0.90
- Context256K
Vision Medium is our pick
Vision Medium is the better all-round choice, scoring 84/100 against Qwen2.5-VL 7B Instruct (40) and Codestral (29). It leads on inputs & features and context window. 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 priceCodestralCodestral $0.45 · Qwen2.5-VL 7B Instruct $0.525 per 1M tokens (3:1 blend) · Vision Medium unpriced
- Longest contextVision MediumVision Medium 1,000,000 · Codestral 256,000 · Qwen2.5-VL 7B Instruct 131,072 tokens
- Widest inputsVision MediumVision Medium: Text, Images, PDFs, Audio, Video · Qwen2.5-VL 7B Instruct: Text, Images · Codestral: Text
- Self-hostingQwen2.5-VL 7B Instruct and CodestralPublishes downloadable weights
| Measure | Weight | Vision Medium | Qwen2.5-VL 7B Instruct | Codestral |
|---|---|---|---|---|
| Inputs & features | 60% | 100 | 50 | 25 |
| Context window | 40% | 60 | 24 | 36 |
| Overall | 100% | 84/100 | 40/100 | 29/100 |
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 | Vision MediumVispark | Qwen2.5-VL 7B InstructAlibaba (Qwen) | CodestralMistral AI |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | — | — | — |
| ECI rank | — | — | — |
| Price per million tokens | |||
| Input | — | $0.35 | $0.30 (best) |
| Output | — | $1.05 | $0.90 (best) |
| Cached input | — | — | $0.03 |
| Blended (3:1) | — | $0.525 | $0.45 (best) |
| Long-context rate | — | Same rate | Same rate |
| Price source | — | Official Alibaba API | Official Mistral API |
| Limits | |||
| Context window | 1,000,000 tokens (best) | 131,072 tokens | 256,000 tokens |
| Max output | 65,536 tokens (best) | 8,192 tokens | 4,096 tokens |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | No |
| PDFs | Yes | No | No |
| Audio | Yes | No | No |
| Video | Yes | No | No |
| Reasoning | Yes | No | No |
| Tool calling | Yes | Yes | Yes |
| Structured output | Yes | No | No |
| Availability | |||
| Weights | Proprietary | Open | Open |
| API model ID | — | qwen2-5-vl-7b-instruct | codestral-latest |
| API providers | — | 1 | 3 (best) |
| Released | May 15, 2024 | Sep 2024 | May 29, 2024 |
| Knowledge cutoff | — | Apr 2024 | Oct 2024 |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
- Vision Medium—
- Qwen2.5-VL 7B Instruct$5.60
- Codestral$4.80
Which should you choose?
Which is better: Vision Medium, Qwen2.5-VL 7B Instruct or Codestral?
Vision Medium is the better all-round choice, scoring 84/100 against Qwen2.5-VL 7B Instruct (40) and Codestral (29). It leads on inputs & features and context window. 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, Vision Medium, Qwen2.5-VL 7B Instruct or Codestral?
Codestral is cheaper at $0.30 input / $0.90 output per million tokens (official Mistral API price). Qwen2.5-VL 7B Instruct costs $0.35 input / $1.05 output per million tokens (official Alibaba API price). At a typical mix of three input tokens to one output token, that is $0.45 per million tokens for Codestral versus $0.525 for Qwen2.5-VL 7B Instruct (1.2× as much). Vision Medium has no published per-token price.
Which scores higher on benchmarks?
There is no independent benchmark that covers all three models yet. Vision Medium has not been scored yet, Qwen2.5-VL 7B Instruct has not been scored yet and Codestral has not been scored yet.
Which is better for coding?
There are no published SWE-bench Verified results for Vision Medium, Qwen2.5-VL 7B Instruct and Codestral 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?
Vision Medium has the largest context window at 1,000,000 tokens, against 256,000 for Codestral and 131,072 for Qwen2.5-VL 7B Instruct. Maximum output per response: Vision Medium up to 65,536, Qwen2.5-VL 7B Instruct up to 8,192, Codestral up to 4,096 tokens.
Which can read images, PDFs, audio or video?
Vision Medium accepts text, images, PDFs, audio and video; Qwen2.5-VL 7B Instruct accepts text and images; Codestral accepts text. Vision Medium handles the widest range of inputs.
Are any of these open source?
Qwen2.5-VL 7B Instruct and Codestral publishes its weights and can be self-hosted; Vision Medium is proprietary.
Which is newer?
Qwen2.5-VL 7B Instruct is the newest, released Sep 2024. Codestral came out May 29, 2024; Vision Medium came out May 15, 2024. Knowledge cutoff: Qwen2.5-VL 7B Instruct Apr 2024, Codestral Oct 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.
