Expired60% confidenceFactExact time
As of 13:45, MiMo V2.6 had trained on 727,000 samples, with 2.6 Pro costing $883,000 (approximately 5.92 million RMB) and 2.6 Flash costing $393,000 (approximately 2.63 million RMB), totaling approximately 8.55 million RMB.
2
Sources
60%
Confidence
Short-term (~14 days)
Relevance
9/18/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/2/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedAt the time of observation, the cumulative training cost for both models was approximately $1.06 million (approximately $741k for Pro and approximately $322k for Flash), and continued to climb.74% similarUnverifiedMeta used more than 16,000 H100 GPUs to train Llama 3.1 405B, with training taking several months and costs potentially exceeding $100 million.66% similarUnverifiedxAI released Grok 4.5, reportedly co-trained with Cursor, priced at $2 per million input tokens and $6 per million output tokens.63% similarUnverifiedIn the 3D art museum test, Opus 4.8 output 437,000 tokens and Fable 5 output 280,000 tokens, with Fable 5 costing approximately $64, which is about 37% more expensive than Opus 4.8's approximately $46.63% similarVerifiedAccounting for real usage scenarios, the actual cost advantage of Opus 5 over Fable is approximately 20% to 25%, with Fable costing about $2.75 per task and Opus 5 about $2.03.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/930924API
curl https://kongchang.com/api/v1/knowledge/claims/930924MCP
get_claim(id=930924)