MDL
Qwen3 14B
Alibaba
Verified against Artificial Analysis · Jul 13, 2026
Capability
54
percentile index
Price /M
$0.35 / $4.2
in / out
Context
131.7K
tokens
Avg score
41.5
11 benchmarks
AA Intelligence
10.4
index
Speed
65 tok/s
TTFT 1.2s
Released
Apr 28, 2025
About
Qwen3 14B is a open-weights model in the Qwen3 family from Alibaba. Benchmarked on 11 evals, averaging 41.5.
Benchmark scores · 11
avg 41.5 — higher is betterThe research behind this model
all papers mentioning it →TRACE: A Unified Rollout Budget Allocation Framework for Efficient Agentic Reinforcement Learning▲ 18 on HF · Jun 9, 2026Verifiable Environments Are LEGO Bricks: Recursive Composition for Reasoning Generalization▲ 7 on HF · Jun 10, 2026ECHO: Terminal Agents Learn World Models for Free▲ 7 on HF · May 23, 2026Simorgh at SemEval-2026 task 7: Region-Aware Hybrid Retrieval for Low-Resource Cultural Reasoning in Multilingual Question AnsweringMay 26, 2026MGRetrieval: Memory-Guided Reflective Retrieval for Long-Term Dialogue AgentsMay 22, 2026
Mentions matched by name in title/abstract — from the arXiv + HF daily corpus.
Performance & price
Output speed65 tok/s
Latency (TTFT)1.2s
First answer32s
Price in / out$0.35 / $4.2 per 1M
Blended price$1.31 / 1M
Coding index13.8
Math index55.7
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$4.2/M