Qwen3 VL 235B A22B Instruct
AlibabaReleased Sep 23, 2025Open weights · Apache 2.0qwen3-vl-235b-a22b
- Reasoning
- no
- Knowledge cutoff
- September 2025
- Context
- 262K tokens
- Max output
- 32k tokens
- Input
- text, image, video
- API features
- tools, structured outputs, JSON mode, seed, logprobs
- Alibaba's API
- yes, as qwen3-vl-235b-a22b-instruct
Read Oct 23, 06:00 UTC from the hosts' published offers and OpenRouter, spec sheet from Alibaba · cutoff: aiknowledgecutoff.com
The cheapest host, DeepInfra, serves fp8 weights (maker: bf16). Weights and context per host
Speed by provider
Output speed of Qwen3 VL 235B A22B Instruct at each host, fastest first
- Venice43 tok/s1.49 s to first token
- Parasail30 tok/s0.86 s to first token
- DeepInfra13 tok/s1.04 s to first token
Speed: OpenRouter, median of the last 30 min, read Oct 23, 06:00 UTC · a snapshot of real traffic before the daily read, not a daily average · bars from zero.
Run it yourself
open weights, Apache 2.0 · fits 1 of the 6 workstations in our guideSelf-host guide ›- Parameters
- 235B
- 22B active per token (MoE)
- Memory needed
- 141 GB
- in INT4 · 282 GB in INT8
- Smallest cloud setup
- 2× H100
- 160 GB · $4.30/h, about $3,139 a month
- Worth it above
- 3B tokens
- a month, against the cheapest API at $1.08 per M
| Quantization | Memory | Smallest cloud GPU | Quality |
|---|---|---|---|
| INT4 (Q4) | 141 GB | 2× H100 · $4.30/h | smallest footprint, minor quality loss |
| INT8 (Q8) | 282 GB | 4× H100 · $8.60/h | near-lossless |
| FP16 (full) | 564 GB | 8× H100 · $17.20/h | reference quality |
At home: in INT4 it fits a Mac Studio M5 Ultra (256 GB unified), about $9,499 once, or $0.3615/h over 3 years, electricity aside; it pays for itself after about 9B tokens against the cheapest API.
Memory ≈ parameters × bytes per parameter × 1.2 · cloud price: on-demand rate, running 24/7 · builds: GGUF, AWQ, MLX, GPTQ, Ollama · break-even calculator
| Provider | Input | Output | Cache ↑ | tok/s | First token | Uptime 24h | Checked |
|---|---|---|---|---|---|---|---|
| Venicefp8 (bf16) · 125K ctx (256K) · 16K out | $0.21 | $1.90 | $0.10 | 43 | 1.49 s | 96.5% | Oct 23 |
| Parasailfp8 (bf16) · 128K ctx (256K) · 32K out | $0.21 | $1.90 | $0.10 | 30 | 0.86 s | 99.7% | Oct 23 |
| DeepInfrafp8 (bf16) · 256K ctx · 16K out | $0.20 | $0.88 | $0.11 | 13 | 1.04 s | 84.8% | Oct 23 |
| Alibabaprecision not read: host terms · 128K ctx (256K) · 32K out | $0.26 | $1.04 | — | — | — | 100% | Oct 9 |
| Novita AIprecision not stated · 128K ctx (256K) · 32K out | $0.30 | $1.50 | — | — | — | — | Oct 8 |
5 of 5 providers · prices per million tokens, read Oct 23, 06:00 UTC · speed: OpenRouter, median of the last 30 min, read Oct 23, 06:00 UTC · uptime: OpenRouter, last 24 h · under each host: weights precision, context and max output as served, the maker's value in brackets when the host's differs
| Provider | Weights | Context | Max output | Tools | Structured outputs | Reasoning |
|---|---|---|---|---|---|---|
| DeepInfra | fp8 | 256K | 16K | yes | yes | no |
| Alibaba | not read: host terms | 128K | 32K | yes | yes | no |
| Novita AI | not stated by the host | 128K | 32K | — | — | — |
| Venice | fp8 | 125K | 16K | yes | yes | no |
| Parasail | fp8 | 128K | 32K | yes | yes | no |
Weights, context and max output: read at each host's own catalogue or docs where its terms allow, otherwise as the host declares to OpenRouter; hover a cell for its source and date · tools, structured outputs, reasoning: as declared to OpenRouter · not verified by us
Qwen3 VL 235B A22B Instruct spec sheet, field by field, with sources
Qwen3 VL 235B A22B Instruct spec sheetmaker-declared
What Alibaba declares about the model itself — read from its own weights repository and documentation, not from any host. What each endpoint actually serves is in the pricing above: the two can differ, and both are true.
- Context window
- 262,144 tokens
- Parameters
- 235,670,022,896≈ 235.7B
- Active experts
- 8 of 128 active per token
- Weight precision
- BF16
- License
- Apache 2.0 — open weights, permissivecurated · read 2026-09-07Scope: the licence on the published weights, and nothing else. It is not the maker's acceptable-use policy, which is a separate document, and it is not the contract you sign with whichever host you call — hosts set their own terms, which this index does not read.
- Knowledge cutoff
- not published by the maker≈ September 2025 third-party estimate — not the maker's figurecurated
- Modalities
- text · image · video
- Max output (official)
- to verify
- Release date
- to verify≈ September 23, 2025 third-party catalogue — not the maker's figure
- Maker lifecycle status
- to verify
Also declared in the weights
- Architecture class
- Qwen3VLMoeForConditionalGeneration
- Layers
- 94
- Vocabulary
- 151,936 tokens
- Weights first published
- September 22, 2025
- Weights last updated
- November 26, 2025
Qwen/Qwen3-VL-235B-A22B-Instruct · config.json @ 710c138 · read 2026-10-08 · 6 of 10 fields published by the maker
| Provider | Chatbot 1,000 in / 500 out | RAG / long context 8,000 in / 500 out | Batch summarize 4,000 in / 1,000 out |
|---|---|---|---|
| DeepInfra | $0.64 | $2.04 | $1.68 |
| Alibaba | $0.78 | $2.60 | $2.08 |
| Novita AI | $1.05 | $3.15 | $2.70 |
| Venice | $1.16 | $2.63 | $2.74 |
| Parasail | $1.16 | $2.63 | $2.74 |
What each host bills for 1,000 requests of a representative job
Your own workload
Other Alibaba models · all its models, hosts and history ›
- Qwen3.8 27B16 providers · from $1.29
- Qwen3.6 35B A3B12 providers · from $1.00
- Qwen3 235B A22B Instruct 250710 providers · from $0.44
- Qwen3.5 397B A17B10 providers · from $3.45
- Qwen3.6 27B9 providers · from $2.25
- Qwen3.5-35B-A3B8 providers · from $1.14