GPT-6 Astra vs. DeepSeek V4 Flash 0731: Vergleich der Veröffentlichungen

Vergleich der Veröffentlichungen GPT-6 Astra und DeepSeek V4 Flash 0731: 6 bzw. 2 Modelle mit unterschiedlichen Eigenschaften bei Intelligenz, Leistung und Preis. Unten werden die wichtigsten Kennzahlen aller 8 Modelle verglichen.

  • Bei der Intelligenz liegt GPT-6 Astra vorn: GPT-6 Astra (max) mit 53, gegenüber DeepSeek V4 Flash Vision (Reasoning, Max Effort) mit 35 bei DeepSeek V4 Flash 0731.
  • Bei der Ausgabegeschwindigkeit ist DeepSeek V4 Flash 0731 am schnellsten: DeepSeek V4 Flash Vision (Reasoning, Max Effort) mit 119 Tokens/s, gegenüber GPT-6 Astra (max) mit 54 Tokens/s bei GPT-6 Astra.
  • Bei den Kosten pro Aufgabe ist DeepSeek V4 Flash 0731 am günstigsten: DeepSeek V4 Flash 0731 (Reasoning, Max Effort) mit $0.22, gegenüber GPT-6 Astra (low) mit $0.82 bei GPT-6 Astra.

Verglichene Veröffentlichungen

OpenAIGPT-6 AstraDeepSeekDeepSeek V4 Flash 0731

Intelligenz

Artificial Analysis Intelligence Index

Artificial Analysis Intelligence Index v4.3 incorporates 10 evaluations: AA-Briefcase, GDPval-AA v2, AutomationBench-AA, Terminal-Bench v4.0, SciCode, Humanity's Last Exam, GDP.pdf, CritPt, AA-Omniscience, AA-LCR v1.1

Artificial Analysis Intelligence Index v4.3 includes: AA-Briefcase, GDPval-AA v2, AutomationBench-AA, Terminal-Bench v4.0, SciCode, Humanity's Last Exam, GDP.pdf, CritPt, AA-Omniscience, AA-LCR v1.1. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

Intelligence Index vs. Cost per Intelligence Index Task

Artificial Analysis Intelligence Index · Weighted average cost (USD) per Artificial Analysis Intelligence Index task
Most attractive quadrant
GPT-6 Astra
DeepSeek V4 Flash 0731
Pareto line

Weighted average cost per Intelligence Index task. Each evaluation’s cost is calculated from input, cache hit, cache write, reasoning, and answer token prices, divided by task count, and weighted by its Intelligence Index weight.

Artificial Analysis Intelligence Index v4.3 includes: AA-Briefcase, GDPval-AA v2, AutomationBench-AA, Terminal-Bench v4.0, SciCode, Humanity's Last Exam, GDP.pdf, CritPt, AA-Omniscience, AA-LCR v1.1. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

Kosten

Cost per Intelligence Index Task

Weighted average cost (USD) per Artificial Analysis Intelligence Index task, segmented by token type. Lower is better

Weighted average cost per Intelligence Index task. Each evaluation’s cost is calculated from input, cache hit, cache write, reasoning, and answer token prices, divided by task count, and weighted by its Intelligence Index weight.

Geschwindigkeit und Latenz

Output Speed

Output tokens per second · Higher is better

Tokens per second received while the model is generating tokens (ie. after first chunk has been received from the API for models which support streaming).

Figures represent performance of the model's first-party API (e.g. OpenAI for o1) or the median across providers where a first-party API is not available (e.g. Meta's Llama models).

Fähigkeitswerte

Fähigkeitsindizes

Misst die Leistung von Modellen bei bestimmten Fähigkeiten und in bestimmten Branchen
Finance & Accounting Index

Incorporates 5 evaluations: AA-Omniscience, GDPval-AA v2, Humanity's Last Exam, 𝜏³-Banking, AA-LCR v1.1 · Higher is better

Strategy & Ops Index

Incorporates 4 evaluations: AA-Omniscience, GDPval-AA v2, 𝜏³-Banking, AA-LCR v1.1 · Higher is better

Legal Index

Incorporates 5 evaluations: AA-Omniscience, GDPval-AA v2, AA-LCR v1.1, 𝜏³-Banking, Humanity's Last Exam · Higher is better

Healthcare & Medical Index

Incorporates 5 evaluations: AA-Omniscience, GDPval-AA v2, MLCR-AA, Humanity's Last Exam, 𝜏³-Banking · Higher is better

Engineering Index

Incorporates 6 evaluations: AA-Omniscience, Humanity's Last Exam, GPQA Diamond, CritPt, GDPval-AA v2, Terminal-Bench v2.1 · Higher is better

Economics Index

Incorporates 4 evaluations: AA-Omniscience, Humanity's Last Exam, GDPval-AA v2, AA-LCR v1.1 · Higher is better

Weitere Details

Gewichte
Anbieter-Benchmarks
GPT-6 Astra (max)
Logo von OpenAIOpenAI
53
-
1M
7,7 $
54
Nicht verfügbar
OpenAI
Anzeigen
GPT-6 Astra (xhigh)
Logo von OpenAIOpenAI
53
-
1M
7,7 $
51
Nicht verfügbar
OpenAI
Anzeigen
GPT-6 Astra (high)
Logo von OpenAIOpenAI
51
-
1M
7,7 $
49
Nicht verfügbar
OpenAI
Anzeigen
GPT-6 Astra (medium)
Logo von OpenAIOpenAI
50
-
1M
7,7 $
44
Nicht verfügbar
OpenAI
Anzeigen
GPT-6 Astra (low)
Logo von OpenAIOpenAI
46
-
1M
7,7 $
47
Nicht verfügbar
OpenAI
Anzeigen
GPT-6 Astra (Non-reasoning)
Logo von OpenAIOpenAI
45
-
1M
7,7 $
-
Nicht verfügbar
OpenAI
Anzeigen
DeepSeek V4 Flash Vision (Reasoning, Max Effort)
Logo von DeepSeekDeepSeek
35
284B
13B während der Inferenz aktiv
1M
0,2 $
119
Nicht verfügbar
DeepSeekFireworks
Anzeigen
DeepSeek V4 Flash 0731 (Reasoning, Max Effort)
Logo von DeepSeekDeepSeek
35
284B
13B während der Inferenz aktiv
1M
0,2 $
118
🤗
CoreWeaveWaferNebius
+12
Anzeigen