Compare AI models

Specifications
Parameters
—2.4T
Context window
1M262k
API pricing
Input price
$5.00—
Output price
$25.00—
Cached input price
$0.50—
Cheapest input
$5.00Amazon Bedrock—
Cheapest output
$25.00Amazon Bedrock—
$2,250Anthropic list price—
Benchmarks
Terminal-Bench 2.1
74.6%76.03%
GPQA Diamond
93.6%92.55%
Overview
CompanyAnthropicQwen
Release dateMay 28 2026Aug 12 2026
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

Claude Opus 4.8 and Qwen3.8-2.4T-A95B are evenly matched across the 2 benchmarks they both report (Terminal-Bench 2.1, GPQA Diamond). Only Claude Opus 4.8 has a verified first-party price for an available API: $5.00 per million input tokens and $25.00 per million output tokens. No currently available pay-as-you-go API rate is tracked for Qwen3.8-2.4T-A95B. Claude Opus 4.8 shipped 76 days before Qwen3.8-2.4T-A95B, so benchmark comparisons should account for the intervening progress.

Context windows are 1M (Claude Opus 4.8) vs 262k (Qwen3.8-2.4T-A95B). Claude Opus 4.8 is closed, while Qwen3.8-2.4T-A95B is open weight.

On Terminal-Bench 2.1, Qwen3.8-2.4T-A95B leads at 76.03% vs Claude Opus 4.8 at 74.6%. On GPQA Diamond, Claude Opus 4.8 leads at 93.6% vs Qwen3.8-2.4T-A95B at 92.55%.