GPT-6 Astra vs. Qwen3.5 397B A17B: comparação de lançamentos

Comparação dos lançamentos GPT-6 Astra e Qwen3.5 397B A17B: 6 e 2 modelos, respectivamente, cada um com diferentes características de inteligência, desempenho e preço. Abaixo, comparamos as principais métricas dos 8 modelos.

  • Em inteligência, GPT-6 Astra tem a pontuação mais alta: GPT-6 Astra (max), com 53, diante de Qwen3.5 397B A17B (Non-reasoning), com 21 de Qwen3.5 397B A17B.
  • Em velocidade de saída, Qwen3.5 397B A17B é o mais rápido: Qwen3.5 397B A17B (Reasoning), com 81 tokens/s, diante de GPT-6 Astra (max), com 54 tokens/s de GPT-6 Astra.
  • Em custo por tarefa, Qwen3.5 397B A17B é o mais barato: Qwen3.5 397B A17B (Reasoning), com $0.47, diante de GPT-6 Astra (low), com $0.82 de GPT-6 Astra.

Lançamentos comparados

OpenAIGPT-6 AstraAlibabaQwen3.5 397B A17B

Inteligência

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
Estimate (independent evaluation forthcoming)

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
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.

Custo

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.

Velocidade e latência

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).

Pontuações de capacidades

Índices de capacidades

Mede o desempenho dos modelos em capacidades e setores específicos
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

Mais detalhes

Pesos
Benchmarks de provedores
GPT-6 Astra (max)
Logo: OpenAIOpenAI
53
-
1M
US$ 7,7
54
Não disponível
OpenAI
Ver
GPT-6 Astra (xhigh)
Logo: OpenAIOpenAI
53
-
1M
US$ 7,7
51
Não disponível
OpenAI
Ver
GPT-6 Astra (high)
Logo: OpenAIOpenAI
51
-
1M
US$ 7,7
49
Não disponível
OpenAI
Ver
GPT-6 Astra (medium)
Logo: OpenAIOpenAI
50
-
1M
US$ 7,7
44
Não disponível
OpenAI
Ver
GPT-6 Astra (low)
Logo: OpenAIOpenAI
46
-
1M
US$ 7,7
47
Não disponível
OpenAI
Ver
GPT-6 Astra (Non-reasoning)
Logo: OpenAIOpenAI
45
-
1M
US$ 7,7
-
Não disponível
OpenAI
Ver
Qwen3.5 397B A17B (Non-reasoning)
Logo: AlibabaAlibaba
21
397B
17B ativos durante a inferência
262k
US$ 0,9
76
🤗
WaferDeepInfraAlibaba Cloud
+3
Ver
Qwen3.5 397B A17B (Reasoning)
Logo: AlibabaAlibaba
19
397B
17B ativos durante a inferência
262k
US$ 0,9
81
🤗
Alibaba CloudWaferParasail
+6
Ver