Expired50% confidenceBenchmarkExact time
Chinese open-source models in the DeepSeek v4 series approach or surpass some closed-source frontier models on coding benchmarks, with inference costs approximately one-tenth that of GPT-4o.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/12/2026
First Seen
Valid until: 10/10/2026(expired)
Sources
Related Claims
UnverifiedChinese open-source models such as DeepSeek V4 continue to achieve breakthroughs in ultra-long context (million-token scale) and cost efficiency.78% similarVerifiedDeepSeek was incubated by quantitative private equity giant High-Flyer, and its R1 series models achieve reasoning capabilities close to GPT-4 level at extremely low training costs.75% similarUnverifiedDeepSeek reaches or surpasses GPT-4-level performance on programming benchmarks such as HumanEval and MBPP, with API pricing approximately 1/20 to 1/30 that of equivalent OpenAI models73% similarUnverifiedDeepSeek V3 has achieved industry-leading levels in code generation capability across multiple programming benchmarks, and its API pricing has a significant cost advantage compared to the GPT series.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/491841API
curl https://kongchang.com/api/v1/knowledge/claims/491841MCP
get_claim(id=491841)