Noometry

Model comparison

DBRX vs GPT-6 Astra

GPT-6 Astra is the stronger model overall, scoring 70.8 to 29.4 on the Noometry Index.

Last verified . 18 shared benchmarks.

DBRX Databricks

29.4

Rank #311 Confirmed

GPT-6 Astra OpenAI

70.8

Rank #1 Confirmed

Summary

  • They share 18 benchmarks with published results for both. DBRX scores higher in 0 categories and GPT-6 Astra in 8 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in math, where GPT-6 Astra leads 93.5 to 24.3.
  • The biggest single-benchmark swing is GPQA Diamond: 32.9% for DBRX and 95.8% for GPT-6 Astra.
  • DBRX has downloadable open weights; the other is API-only.

Side by side

DBRXGPT-6 Astra
ProviderDatabricksOpenAI
Noometry Index29.470.8
Released2024-03-272026-09-03
WeightsOpenProprietary
Context window—1.05M
Max output—128K
Input $ / M tokens—$10
Output $ / M tokens—$50
Results tracked2156

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

Category by category

Coding GPT-6 Astra leads

DBRX: 32.9 (#266), GPT-6 Astra: 73.7 (#2)

BenchmarkDBRXGPT-6 Astra
LMArena Coding11321487
DeepSWE—74.1%
FrontierCode—53.3%
LMArena WebDev—1786
FrontierSWE—65.5%
SciCode—56.5%
GSO—79.4%
WeirdML—93.6%
MirrorCode—46.7%
ALE-Bench—2,951
HumanEval+70.1%—
MBPP+55.8%—

Agentic & Tool Use Not comparable

DBRX: —, GPT-6 Astra: 52.9 (#3)

BenchmarkDBRXGPT-6 Astra
APEX-Agents—64.7%
Remote Labor Index—20.8%
BALROG—68.3%
GDP.pdf—34.2%
Vending-Bench 2—15,515

Reasoning GPT-6 Astra leads

DBRX: 21.4 (#222), GPT-6 Astra: 85.1 (#1)

BenchmarkDBRXGPT-6 Astra
LMArena Hard Prompts11131462
ARC-AGI-2—95%
NYT Connections (extended)—98.1%
ARC-AGI-1—98.5%
CritPt—31.7%
Chess Puzzles—72%
EBR-Bench—76.2%
Mystery Game Puzzles—84%
DTBench—97.3%
LMCA—64.4%
Bench to the Future 3—0.14
Epoch Capabilities Index—166.45

Math GPT-6 Astra leads

DBRX: 24.3 (#269), GPT-6 Astra: 93.5 (#2)

BenchmarkDBRXGPT-6 Astra
LMArena Math11451465
FrontierMath (Tiers 1-3)—93.7%
FrontierMath Tier 4—97.6%
OTIS Mock AIME 2024-2025—100%
ProofBench—99%
MATH Level 511.7%—
FrontierMath Erdős—2.9%

Knowledge GPT-6 Astra leads

DBRX: 14.9 (#294), GPT-6 Astra: 75.3 (#1)

BenchmarkDBRXGPT-6 Astra
GPQA Diamond32.9%95.8%
LMArena Expert10761483
Humanity's Last Exam—54.8%
SimpleQA Verified—75.6%
Vectara Hallucination Rate—8.7%

Multimodal Not comparable

DBRX: —, GPT-6 Astra: 55.0 (#3)

BenchmarkDBRXGPT-6 Astra
LMArena Vision—1281
Blueprint-Bench 2—49.7%
Furniture Assembly—80%
LMArena Document—1468

Multilingual GPT-6 Astra leads

DBRX: 29.3 (#268), GPT-6 Astra: 53.7 (#61)

BenchmarkDBRXGPT-6 Astra
LMArena Non-English10711430
LMArena Chinese10681484
LMArena French10961456
LMArena German10571440
LMArena Japanese9901379
LMArena Korean9931426
LMArena Russian10781436
LMArena Spanish10641407

Instruction Following GPT-6 Astra leads

DBRX: 57.5 (#270), GPT-6 Astra: 76.3 (#44)

BenchmarkDBRXGPT-6 Astra
LMArena Instruction Following11121450

Long Context GPT-6 Astra leads

DBRX: 33.7 (#258), GPT-6 Astra: 44.5 (#62)

BenchmarkDBRXGPT-6 Astra
LMArena Longer Query11121456

Writing & Preference GPT-6 Astra leads

DBRX: 33.6 (#275), GPT-6 Astra: 75.3 (#7)

BenchmarkDBRXGPT-6 Astra
LMArena Text11191441
LMArena Creative Writing11041418
LMArena Multi-Turn11111448
EQ-Bench Creative Writing—2173

Frequently asked questions

Is DBRX better than GPT-6 Astra?

GPT-6 Astra is the stronger model overall, scoring 70.8 to 29.4 on the Noometry Index.

Is DBRX or GPT-6 Astra better for coding?

GPT-6 Astra scores higher on coding benchmarks: 73.7 versus 32.9 in the Noometry coding category.

How many benchmarks do DBRX and GPT-6 Astra share?

18 benchmarks have published results for both models. DBRX has 21 scored results on Noometry and GPT-6 Astra has 56.

Related comparisons

Go deeper