Neither wins outright: Mistral: Mistral Medium 3.1 (batch) is the pick when you need file / PDF input, Qwen: Qwen3 VL 32B Instruct when cost matters at scale.
| Spec | Mistral: Mistral Medium 3.1 (batch) | Qwen: Qwen3 VL 32B Instruct |
|---|---|---|
| Provider | Mistral AI | Qwen |
| Input price (per 1M tokens) | $0.20 | $0.10 |
| Output price (per 1M tokens) | $1.00 | $0.42 |
| Blended price (3:1 input/output) | $0.40 | $0.18 |
| Cheaper than … of all models we track | 58% | 70% |
| Context window | 131K tokens (~197 pages) | 131K tokens (~197 pages) |
| Larger context than … of all models | 23% | 23% |
| Max output per response | 105K tokens (~78,600 words) | 33K tokens (~24,600 words) |
| Input types | text, image, file | text, image |
| Output types | text | text |
| Free variant | No | No |
| Listed since | 2025-08-13 | 2025-10-23 |
Monthly cost at list price for five common workloads. Token counts per unit are stated so you can scale them to your own traffic. A workload that does not fit a model's context window is marked rather than priced.
| Workload (per month) | Mistral: Mistral Medium 3.1 (batch) | Qwen: Qwen3 VL 32B Instruct | Difference |
|---|---|---|---|
| Customer-support chatbot 50,000 conversations — each a 6-turn conversation (1,500 in / 400 out tokens) | $35.00 | $16.12 | Qwen: Qwen3 VL 32B Instruct saves $18.88 |
| RAG search over documents 20,000 questions — each a question answered from ~8 retrieved passages (6,000 in / 500 out tokens) | $34.00 | $16.64 | Qwen: Qwen3 VL 32B Instruct saves $17.36 |
| Coding / tool-using agent 2,000 tasks — each a multi-step task with tool calls and file context (60,000 in / 6,000 out tokens) | $36.00 | $17.47 | Qwen: Qwen3 VL 32B Instruct saves $18.53 |
| Batch summarisation 10,000 documents — each a 6,000-word report summarised to a page (8,000 in / 600 out tokens) | $22.00 | $10.82 | Qwen: Qwen3 VL 32B Instruct saves $11.18 |
| Long-document analysis 500 documents — each a 150-page contract or codebase read in one pass (100,000 in / 2,000 out tokens) | $11.00 | $5.62 | Qwen: Qwen3 VL 32B Instruct saves $5.38 |
| Capability | Mistral: Mistral Medium 3.1 (batch) | Qwen: Qwen3 VL 32B Instruct | What it lets you do |
|---|---|---|---|
| Tool / function calling | ✓ Yes | ✓ Yes | call your APIs and run agent loops |
| Structured output (JSON schema) | ✓ Yes | ✓ Yes | return JSON that validates against your schema |
| Reasoning / thinking mode | — | — | spend extra tokens thinking before answering hard problems |
| Image input | ✓ Yes | ✓ Yes | read screenshots, charts and scanned pages |
| Audio input | — | — | take speech or audio directly |
| File / PDF input | ✓ Yes | — | accept documents without your own parsing step |
| Image output | — | — | generate images, not just text |
| Built-in web search | — | — | answer from live web results |
| Seed (reproducible sampling) | ✓ Yes | ✓ Yes | repeat a generation for tests and evals |
| Log-probabilities | — | ✓ Yes | read token confidence for classification and scoring |
From each model's published parameters and input/output types. “—” means not advertised, not proven absent.
| Date | Model | Input / 1M | Output / 1M |
|---|---|---|---|
| 2026-09-10 | Mistral: Mistral Medium 3.1 (batch) | $0.20 | $1.00 |
Neither wins outright: Mistral: Mistral Medium 3.1 (batch) is the pick when you need file / PDF input, Qwen: Qwen3 VL 32B Instruct when cost matters at scale. In short: Mistral: Mistral Medium 3.1 (batch) when you need file / PDF input, you generate long outputs in one go; Qwen: Qwen3 VL 32B Instruct when cost matters at scale, you need log-probabilities.
Qwen: Qwen3 VL 32B Instruct is cheaper: $0.18 vs $0.40 per 1M tokens at a typical 3:1 input/output mix, 55% less. For 50,000 support conversations a month that is $16.12 against $35.00.
Mistral: Mistral Medium 3.1 (batch) costs $0.20 per 1M input tokens and $1.00 per 1M output tokens; Qwen: Qwen3 VL 32B Instruct costs $0.10 input and $0.42 output. Output tokens are billed separately, so long answers shift the gap.
Both take 131K tokens per request — roughly 197 pages of text.
Tool calling — Mistral: Mistral Medium 3.1 (batch): yes, Qwen: Qwen3 VL 32B Instruct: yes. JSON-schema structured output — Mistral: Mistral Medium 3.1 (batch): yes, Qwen: Qwen3 VL 32B Instruct: yes.
Yes, both accept image input.
No free variant is listed for either; both are pay-per-token.
Mistral: Mistral Medium 3.1 (batch) can write up to 105K tokens (about 78,600 words) per response; Qwen: Qwen3 VL 32B Instruct up to 33K (about 24,600 words).
Mistral: Mistral Medium 3.1 (batch) was listed on 2025-08-13; Qwen: Qwen3 VL 32B Instruct on 2025-10-23.
Mistral: Mistral Medium 3.1 (batch): input price cut 50% ($0.40 → $0.20/1M), output price cut 50% ($2.00 → $1.00/1M) on 2026-09-10. Qwen: Qwen3 VL 32B Instruct: price unchanged since we started tracking it on 2026-03-10.
Mistral Medium 3.1 is an updated version of Mistral Medium 3, which is a high-performance enterprise-grade language model designed to deliver frontier-level capabilities at significantly reduced operational cost. It balances...
Full Mistral: Mistral Medium 3.1 (batch) specs →Qwen3-VL-32B-Instruct is a large-scale multimodal vision-language model designed for high-precision understanding and reasoning across text, images, and video. With 32 billion parameters, it combines deep visual perception with advanced text...
Full Qwen: Qwen3 VL 32B Instruct specs →