Unverified50% confidenceBenchmarkExact time
Early versions of the Claude series already supported a context of 100K tokens, far exceeding the 8K/32K limit of GPT-4 at the time.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/5/2026
First Seen
Valid until: 12/4/2026
Sources
Related Entities
Related Claims
Unverified早期Codex的上下文窗口约为4096个token，对应大约150-300行代码；GPT-4系列将这一限制扩展至128K token67% similarUnverifiedFor a typical 7B model (32 layers, 32 heads, 128 dimensions) processing a 2048-token sequence at FP16 precision with a batch size of 8, the KV Cache consumes approximately 4 GB of VRAM.64% similarUnverifiedThe 200K context of the Claude 3 series can accommodate approximately 15,000 lines of code.63% similarUnverifiedOn a single 32GB GPU under q4_0 quantization, the context length increased from 82,432 tokens before the update to 199,680 tokens after the update.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/862213API
curl https://kongchang.com/api/v1/knowledge/claims/862213MCP
get_claim(id=862213)