Noometry

Model comparison

GPT-6 Astra vs Qwen3-4B

GPT-6 Astra is the stronger model overall, scoring 70.8 to 31.9 on the Noometry Index.

Last verified . 4 shared benchmarks.

GPT-6 Astra OpenAI

70.8

Rank #1 Confirmed

Qwen3-4B Alibaba (Qwen)

31.9

Rank #264 Confirmed

Summary

  • They share 4 benchmarks with published results for both. GPT-6 Astra scores higher in 4 categories and Qwen3-4B in 0 categories; 4 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where GPT-6 Astra leads 85.1 to 19.2.
  • The biggest single-benchmark swing is Chess Puzzles: 72% for GPT-6 Astra and 4% for Qwen3-4B.
  • Qwen3-4B has downloadable open weights; the other is API-only.

Side by side

GPT-6 AstraQwen3-4B
ProviderOpenAIAlibaba (Qwen)
Noometry Index70.831.9
Released2026-09-032025-04-29
WeightsProprietaryOpen
Context window1.05M—
Max output128K—
Input $ / M tokens$10—
Output $ / M tokens$50—
Results tracked566

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Not comparable

GPT-6 Astra: 73.7 (#2), Qwen3-4B: —

BenchmarkGPT-6 AstraQwen3-4B
DeepSWE74.1%—
FrontierCode53.3%—
LMArena WebDev1786—
FrontierSWE65.5%—
SciCode56.5%—
GSO79.4%—
WeirdML93.6%—
LMArena Coding1487—
MirrorCode46.7%—
ALE-Bench2,951—

Agentic & Tool Use GPT-6 Astra leads

GPT-6 Astra: 52.9 (#3), Qwen3-4B: 27.6 (#100)

BenchmarkGPT-6 AstraQwen3-4B
APEX-Agents64.7%—
Berkeley Function Calling Leaderboard—35.7%
Remote Labor Index20.8%—
BALROG68.3%—
GDP.pdf34.2%—
Vending-Bench 215,515—

Reasoning GPT-6 Astra leads

GPT-6 Astra: 85.1 (#1), Qwen3-4B: 19.2 (#268)

BenchmarkGPT-6 AstraQwen3-4B
Chess Puzzles72%4%
ARC-AGI-295%—
NYT Connections (extended)98.1%—
ARC-AGI-198.5%—
CritPt31.7%—
EBR-Bench76.2%—
LMArena Hard Prompts1462—
Mystery Game Puzzles84%—
DTBench97.3%—
LMCA64.4%—
Bench to the Future 30.14—
Epoch Capabilities Index166.45—

Math GPT-6 Astra leads

GPT-6 Astra: 93.5 (#2), Qwen3-4B: 29.7 (#240)

BenchmarkGPT-6 AstraQwen3-4B
OTIS Mock AIME 2024-2025100%52.2%
FrontierMath (Tiers 1-3)93.7%—
FrontierMath Tier 497.6%—
MathArena Final-Answer Competitions—38.5%
ProofBench99%—
LMArena Math1465—
FrontierMath Erdős2.9%—

Knowledge GPT-6 Astra leads

GPT-6 Astra: 75.3 (#1), Qwen3-4B: 33.0 (#208)

BenchmarkGPT-6 AstraQwen3-4B
GPQA Diamond95.8%52.3%
Vectara Hallucination Rate8.7%5.7%
Humanity's Last Exam54.8%—
SimpleQA Verified75.6%—
LMArena Expert1483—

Multimodal Not comparable

GPT-6 Astra: 55.0 (#3), Qwen3-4B: —

BenchmarkGPT-6 AstraQwen3-4B
LMArena Vision1281—
Blueprint-Bench 249.7%—
Furniture Assembly80%—
LMArena Document1468—

Multilingual Not comparable

GPT-6 Astra: 53.7 (#61), Qwen3-4B: —

BenchmarkGPT-6 AstraQwen3-4B
LMArena Non-English1430—
LMArena Chinese1484—
LMArena French1456—
LMArena German1440—
LMArena Japanese1379—
LMArena Korean1426—
LMArena Russian1436—
LMArena Spanish1407—

Instruction Following Not comparable

GPT-6 Astra: 76.3 (#44), Qwen3-4B: —

BenchmarkGPT-6 AstraQwen3-4B
LMArena Instruction Following1450—

Long Context Not comparable

GPT-6 Astra: 44.5 (#62), Qwen3-4B: —

BenchmarkGPT-6 AstraQwen3-4B
LMArena Longer Query1456—

Writing & Preference Not comparable

GPT-6 Astra: 75.3 (#7), Qwen3-4B: —

BenchmarkGPT-6 AstraQwen3-4B
LMArena Text1441—
LMArena Creative Writing1418—
EQ-Bench Creative Writing2173—
LMArena Multi-Turn1448—

Frequently asked questions

Is GPT-6 Astra better than Qwen3-4B?

GPT-6 Astra is the stronger model overall, scoring 70.8 to 31.9 on the Noometry Index.

How many benchmarks do GPT-6 Astra and Qwen3-4B share?

4 benchmarks have published results for both models. GPT-6 Astra has 56 scored results on Noometry and Qwen3-4B has 6.

Related comparisons

Go deeper