MDL
Qwen3 235B A22B 2507 Instruct
Alibaba
Verified against Artificial Analysis · Jul 13, 2026
Capability
68
percentile index
Price /M
$0.7 / $2.8
in / out
Context
—
Avg score
50.2
11 benchmarks
AA Intelligence
18.2
index
Speed
66 tok/s
TTFT 1.2s
Released
Jul 21, 2025
About
Qwen3 235B A22B 2507 Instruct is a open-weights model in the Qwen3 family from Alibaba. Benchmarked on 11 evals, averaging 50.2.
Benchmark scores · 11
avg 50.2 — higher is betterThe research behind this model
all papers mentioning it →ACC: Compiling Agent Trajectories for Long-Context TrainingMay 21, 2026Draft Less, Retrieve More: Hybrid Tree Construction for Speculative DecodingMay 19, 2026GoLongRL: Capability-Oriented Long Context Reinforcement Learning with Multitask AlignmentMay 19, 2026FORGE: Self-Evolving Agent Memory With No Weight Updates via Population BroadcastMay 15, 2026Persona-Model Collapse in Emergent MisalignmentMay 13, 2026
Mentions matched by name in title/abstract — from the arXiv + HF daily corpus.
Performance & price
Output speed66 tok/s
Latency (TTFT)1.2s
Price in / out$0.7 / $2.8 per 1M
Blended price$1.23 / 1M
Math index71.7
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