MDL
Qwen3 Omni 30B A3B Instruct
Alibaba
Verified against Artificial Analysis · Jul 13, 2026
Capability
42
percentile index
Price /M
$0.25 / $0.97
in / out
Context
—
Avg score
30.2
10 benchmarks
AA Intelligence
5.1
index
Speed
109 tok/s
TTFT 1s
Released
Sep 22, 2025
About
Qwen3 Omni 30B A3B Instruct is a open-weights model in the Qwen3 family from Alibaba. Benchmarked on 10 evals, averaging 30.2.
Benchmark scores · 10
avg 30.2 — higher is betterThe research behind this model
all papers mentioning it →OmniVideo-100K: A Dataset for Audio-Visual Reasoning through Structured Scripts and Evidence Chains▲ 31 on HF · Jun 12, 2026Data-Efficient On-Policy Distillation for Automatic Speech RecognitionMay 27, 2026From Text to Voice: A Reproducible and Verifiable Framework for Evaluating Tool Calling LLM AgentsMay 14, 2026Boosting Omni-Modal Language Models: Staged Post-Training with Visually Debiased EvaluationMay 12, 2026Sparse Tokens Suffice: Jailbreaking Audio Language Models via Token-Aware Gradient OptimizationMay 6, 2026
Mentions matched by name in title/abstract — from the arXiv + HF daily corpus.
Performance & price
Output speed109 tok/s
Latency (TTFT)1s
Price in / out$0.25 / $0.97 per 1M
Blended price$0.43 / 1M
Math index52.3
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$0.97/M