The Known Good Updated 2 Sep 2026 Subscribe
DeepSeek: DeepSeek V3 vs QwQ-32B ·DeepSeek vs Alibaba

DeepSeek: DeepSeek V3 vs QwQ-32B — head-to-head comparison

DeepSeek: DeepSeek V3 profile QwQ-32B profile All comparisons
DeepSeek: DeepSeek V3 — DeepSeek
Quality
#43 of 65
62based on 4 of 7 evaluations
Known Good Index
Speed
#314 of 339
17.8
Output tokens per second
Price
#149 of 439
$0.45
Blended per 1M tokens
In $0.26 · Out $1.03
Cache hit price
#59 of 322
$0.020
per 1M tokens (−92%)
Context
#334 of 473
164k
tokens (~300 pages)
QwQ-32B — Alibaba
Quality
#29 of 65
69based on 3 of 7 evaluations
Known Good Index
Speed
unranked
—
Output tokens per second
Price
unranked
—
Blended per 1M tokens
Cache hit price
unranked
—
per 1M tokens
Context
unranked
—
tokens
Summary

On the Known Good Index, DeepSeek: DeepSeek V3 scores 62 and QwQ-32B scores 69 — a 7-point lead for QwQ-32B. They share 9 published evaluations, with DeepSeek: DeepSeek V3 ahead on 1 and QwQ-32B ahead on 8.

Published pricing is tracked for DeepSeek: DeepSeek V3 only, at $0.45 per 1M tokens blended.

DeepSeek: DeepSeek V3 is open weights under Open weights (unrestricted); QwQ-32B is open weights under Open weights (unrestricted). 3 tracked hosts serve one or both.

Generated from tracked data on 3 Sep 2026 · how these comparisons are computed →

At a glance
MetricDeepSeek: DeepSeek V3QwQ-32B
Known Good Index 62 based on 4 of 7 evaluations 69 based on 3 of 7 evaluations
Blended $/1M $0.45 —
Output tok/s 18 —
Context window 164k —
Providers 3 0
Jump to section

Quality

Known Good Index and the capability indices, side by side

Index scores

Each index is min-max normalised across tracked models and scaled 0–100 from published evaluations. We do not run these evaluations.

MetricDeepSeek: DeepSeek V3QwQ-32B
Known Good Index 62 69
Mathematics Index 57 —
Reasoning Index 62 75
Openness Index 60 60
The Known Good

Quality Breakdown

Every evaluation both models report a published score for

Evaluations both models report

Bar = DeepSeek: DeepSeek V3 · marker = QwQ-32B · scores ingested from Epoch AI (CC-BY 4.0), LiveBench (Apache-2.0)

GPQA Diamond71 / 65
Mock AIME 2024–2549 / 59
LiveBench60 / 72
LiveBench Coding62 / 72
LiveBench Data Analysis61 / 65
LiveBench Instruction Following75 / 82
LiveBench Language47 / 51
LiveBench Mathematics61 / 78
LiveBench Reasoning57 / 84

Right-hand figures read DeepSeek: DeepSeek V3 / QwQ-32B. DeepSeek: DeepSeek V3 leads on 1 of 9, QwQ-32B on 8.

The Known Good

Arena Elo

Human preference rating from blind pairwise votes

Arena ratings

Elo from blind pairwise votes, ingested from Arena. Higher is better.

MetricDeepSeek: DeepSeek V3QwQ-32B
Text Arena 1,332 1,329
Text Arena (style controlled) 1,358 1,336
The Known Good

Price & Cost

Published API pricing by token type

Published pricing

USD per 1M tokens. Blended is a 3:1 input:output mix. Live from OpenRouter, cross-checked against models.dev. Lower is better.

MetricDeepSeek: DeepSeek V3QwQ-32B
Input $/1M $0.26 —
Output $/1M $1.03 —
Blended $/1M $0.45 —
Cache read $/1M $0.020 —
The Known Good
We publish blended price rather than cost-per-task. Cost-per-task requires token counts from an evaluation suite we do not run. Blended price is computed from published rates and is fully reproducible — see the pricing methodology.

Context Window

Maximum input accepted and maximum output emitted

Context and output limits

Provider-declared limits, from metadata via models.dev and OpenRouter. Higher is better.

MetricDeepSeek: DeepSeek V3QwQ-32B
Context window 164k —
Max output tokens 16k —
The Known Good

Speed & Latency

Mean throughput and time to first token across hosting providers, over successful measurements from the last 48 hours

Throughput and latency

Means across all tracked providers over successful measurements from the last 48 hours, derived from OpenRouter endpoint telemetry.

MetricDeepSeek: DeepSeek V3QwQ-32B
Output tokens/s 18 —
Time to first token 1.8s —
The Known Good

Providers

Which hosts serve each model, and at what price and speed

3 hosts
Full provider benchmarks →
Provider DeepSeek: DeepSeek V3 blendedDeepSeek: DeepSeek V3 tok/s QwQ-32B blendedQwQ-32B tok/s
StreamLake $0.45 34 — —
DeepInfra $0.46 22 — —
Novita $0.62 17 — —

Per-host pricing and telemetry from OpenRouter. A dash means that host does not serve that model.

The Known Good
Sources for this page
Pricing & providers — OpenRouter Metadata — models.dev Benchmarks — Epoch AI (CC-BY 4.0)Benchmarks — LiveBench (Apache-2.0)Elo — Arena Full attribution →