MDL
Qwen3 VL 235B A22B Instruct
Alibaba
Verified against Artificial Analysis · Jul 13, 2026
Capability
62
percentile index
Price /M
$0.7 / $2.8
in / out
Context
262.1K
tokens
Avg score
44.2
10 benchmarks
AA Intelligence
14.3
index
Speed
57 tok/s
TTFT 1.1s
Released
Sep 23, 2025
About
Qwen3 VL 235B A22B Instruct is a open-weights model in the Qwen3 family from Alibaba. Benchmarked on 10 evals, averaging 44.2.
Benchmark scores · 10
avg 44.2 — higher is betterThe research behind this model
all papers mentioning it →Agent Explorative Policy Optimization for Multimodal Agentic Reasoning▲ 90 on HF · May 27, 2026MobileGym: A Verifiable and Highly Parallel Simulation Platform for Mobile GUI Agent Research▲ 64 on HF · May 25, 2026LabVLA: Grounding Vision-Language-Action Models in Scientific Laboratories▲ 54 on HF · Jun 11, 2026MobileForge: Annotation-Free Adaptation for Mobile GUI Agents with Hierarchical Feedback-Guided Policy Optimization▲ 43 on HF · Jun 18, 2026S-Agent: Spatial Tool-Use Elicits Reasoning for Spatial Intelligence▲ 39 on HF · Jun 18, 2026
Mentions matched by name in title/abstract — from the arXiv + HF daily corpus.
Performance & price
Output speed57 tok/s
Latency (TTFT)1.1s
Price in / out$0.7 / $2.8 per 1M
Blended price$1.23 / 1M
Math index70.7
Median across API providers, via Artificial Analysis. Blended = 3:1 input:output. For reasoning models, latency includes thinking time. Methodology →
History
full ledger →Capability over time
Append-only ledger — every observed change to this model's numbers.
API providers
Artificial Analysis$2.8/M