Qwen3 VL 235B A22B Instruct
by Alibaba (Qwen) · qwen/qwen3-vl-235b-a22b-instruct · #142 cheapest of 325 paid models
Prices updated Jul 22, 2026, 2:53 AM UTC · refreshed hourly
Qwen3-VL-235B-A22B Instruct is an open-weight multimodal model that unifies strong text generation with visual understanding across images and video. The Instruct model targets general vision-language use (VQA, document parsing, chart/table...
Compare vs:Claude Sonnet 5Claude Opus 4.6Claude Haiku 4.5GPT-5.6 Sol
Pricing
Input
$0.21
per 1M tokens
Output
$1.90
per 1M tokens
Blended (3:1)
$0.632
3 input : 1 output
Cache read
$0.10
per 1M cached tokens
Cache write
—
per 1M tokens
Cache & batch economics
Effective prices for Qwen3 VL 235B A22B Instruct with prompt caching.
| Cache hit rate | Effective blended $/1M | Example request* | vs no cache |
|---|---|---|---|
| 0% | $0.632 | $0.0233 | — |
| 50% | $0.591 | $0.0178 | −24% |
| 90% | $0.558 | $0.0134 | −42% |
*Example: 100K sticky context + 2K new input + 1K output. Batch discount is an approximate 50% off list rates (available for OpenAI / Anthropic / Google — not flagged for this creator). Confirm on the vendor’s pricing page.
What a request costs
| Workload | Input tokens | Output tokens | Cost / request | Cost / 1K requests |
|---|---|---|---|---|
| Short chat message | 500 | 300 | $0.000675 | $0.675 |
| Document summary | 8,000 | 1,000 | $0.00358 | $3.58 |
| Codebase question (RAG) | 30,000 | 2,000 | $0.0101 | $10.10 |
| Long-context analysis | 150,000 | 5,000 | $0.041 | $41.00 |
Performance
Intelligence Index
14.3
AA composite quality
Coding Index
—
Math Index
70.7
Agentic Index
—
Output speed
48 tok/s
median
Time to first token
1.11s
median
Time to first answer
1.11s
after reasoning tokens
Artificial Analysis benchmarks
| Benchmark | Score |
|---|---|
| MMLU-Pro | 82.3% |
| GPQA Diamond | 71.2% |
| Humanity's Last Exam | 6.3% |
| LiveCodeBench | 59.4% |
| SciCode | 35.9% |
| AIME 2025 | 70.7% |
| IFBench | 42.7% |
| AA-LCR (Long Context Reasoning) | 31.7% |
| Terminal-Bench Hard | 6.8% |
| τ²-Bench (Telecom) | 35.1% |
Specs
Context window
262K
tokens
Max output
33K
tokens
Input modalities
text, image
Tokenizer
Qwen3