[ { "component_id": "llama2-published-training", "family": "Llama 2", "organization": "Meta", "unit": "million_A100_hours", "quantity": 3.3, "quantity_provenance": "Published hardware-hours; rental price is assumed", "scope": "All released sizes; no generic allowance.", "sources": [ "https://github.com/meta-llama/llama/blob/main/MODEL_CARD.md" ], "conversion_rates": [ 1, 2, 3 ], "low_musd": 3.3, "central_musd": 6.6, "high_musd": 9.899999999999999 }, { "component_id": "llama31-cumulative", "family": "Llama 3 / 3.1 / 3.2 / 3.3", "organization": "Meta", "unit": "million_H100_hours", "quantity": 39.3, "quantity_provenance": "Published hardware-hours; rental price is assumed", "scope": "3.1 reported total;3.0 overlap unresolved and excluded here;3.3 repetition excluded.", "sources": [ "https://github.com/meta-llama/llama-models/blob/main/models/llama3_1/MODEL_CARD.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3_2/MODEL_CARD.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3_2/MODEL_CARD_VISION.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3/MODEL_CARD.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3_3/MODEL_CARD.md" ], "conversion_rates": [ 1.5, 2.5, 4 ], "low_musd": 58.949999999999996, "central_musd": 98.25, "high_musd": 157.2 }, { "component_id": "llama32-text-increment", "family": "Llama 3 / 3.1 / 3.2 / 3.3", "organization": "Meta", "unit": "million_H100_hours", "quantity": 0.916, "quantity_provenance": "Published hardware-hours; rental price is assumed", "scope": "Published text total incl86k logit generation; text/table rounding discrepancy retained in note.", "sources": [ "https://github.com/meta-llama/llama-models/blob/main/models/llama3_1/MODEL_CARD.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3_2/MODEL_CARD.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3_2/MODEL_CARD_VISION.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3/MODEL_CARD.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3_3/MODEL_CARD.md" ], "conversion_rates": [ 1.5, 2.5, 4 ], "low_musd": 1.374, "central_musd": 2.29, "high_musd": 3.664 }, { "component_id": "llama32-vision-increment", "family": "Llama 3 / 3.1 / 3.2 / 3.3", "organization": "Meta", "unit": "million_H100_hours", "quantity": 2.02, "quantity_provenance": "Published hardware-hours; rental price is assumed", "scope": "11B/90B vision additional hours; no second Llama3 base.", "sources": [ "https://github.com/meta-llama/llama-models/blob/main/models/llama3_1/MODEL_CARD.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3_2/MODEL_CARD.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3_2/MODEL_CARD_VISION.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3/MODEL_CARD.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3_3/MODEL_CARD.md" ], "conversion_rates": [ 1.5, 2.5, 4 ], "low_musd": 3.0300000000000002, "central_musd": 5.05, "high_musd": 8.08 }, { "component_id": "llama4-pretraining", "family": "Llama 4 Scout + Maverick", "organization": "Meta", "unit": "million_H100_hours", "quantity": 7.38, "quantity_provenance": "Published hardware-hours; rental price is assumed", "scope": "Published pretraining only; posttraining unknown.", "sources": [ "https://github.com/meta-llama/llama-models/blob/main/models/llama4/MODEL_CARD.md" ], "conversion_rates": [ 1.5, 2.5, 4 ], "low_musd": 11.07, "central_musd": 18.45, "high_musd": 29.52 }, { "component_id": "qwen3-original-eight-s1", "family": "Qwen 3", "organization": "Alibaba", "unit": "1e24_FLOP", "quantity": 15.353999999999997, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "All eight models >30T S1; conservative30T slice, not a claim that complete training is30T.", "sources": [ "https://huggingface.co/Qwen/Qwen3-235B-A22B", "https://arxiv.org/html/2505.09388v1", "https://huggingface.co/Qwen/Qwen3-Coder-480B-A35B-Instruct", "https://qwenlm.github.io/blog/qwen3-coder/", "https://huggingface.co/Qwen/Qwen3-Next-80B-A3B-Instruct" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 12.283199999999999, "central_musd": 27.637199999999996, "high_musd": 61.41599999999999 }, { "component_id": "qwen3-coder-480b", "family": "Qwen 3", "organization": "Alibaba", "unit": "1e24_FLOP", "quantity": 1.575, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "35B active×7.5T tokens; official blog confirmed.", "sources": [ "https://qwenlm.github.io/blog/qwen3-coder/" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 1.26, "central_musd": 2.835, "high_musd": 6.3 }, { "component_id": "qwen3-next-80b", "family": "Qwen 3", "organization": "Alibaba", "unit": "1e24_FLOP", "quantity": 0.27, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "3B active×15T; official card.", "sources": [ "https://huggingface.co/Qwen/Qwen3-Next-80B-A3B-Instruct" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 0.21600000000000003, "central_musd": 0.48600000000000004, "high_musd": 1.08 }, { "component_id": "deepseek-v2", "family": "DeepSeek V2", "organization": "DeepSeek", "unit": "1e24_FLOP", "quantity": 1.0206, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "21B active×8.1T.", "sources": [ "https://huggingface.co/deepseek-ai/DeepSeek-V2" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 0.81648, "central_musd": 1.83708, "high_musd": 4.0824 }, { "component_id": "deepseek-v3-published", "family": "DeepSeek V3 / R1 / V3.x", "organization": "DeepSeek", "unit": "million_H800_hours", "quantity": 2.788, "quantity_provenance": "Published hardware-hours; rental price is assumed", "scope": "Includes published context/posttraining; no inferred R1 or V3.x budget.", "sources": [ "https://huggingface.co/deepseek-ai/DeepSeek-V3-Base", "https://arxiv.org/html/2412.19437v2" ], "conversion_rates": [ 1, 2, 3 ], "low_musd": 2.788, "central_musd": 5.576, "high_musd": 8.363999999999999 }, { "component_id": "deepseek-v4-pro", "family": "DeepSeek V4 Pro + Flash", "organization": "DeepSeek", "unit": "1e24_FLOP", "quantity": 9.702, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "49B active×33T.", "sources": [ "https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro", "https://arxiv.org/html/2606.19348" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 7.7616000000000005, "central_musd": 17.4636, "high_musd": 38.808 }, { "component_id": "deepseek-v4-flash", "family": "DeepSeek V4 Pro + Flash", "organization": "DeepSeek", "unit": "1e24_FLOP", "quantity": 2.496, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "13B active×32T.", "sources": [ "https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro", "https://arxiv.org/html/2606.19348" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 1.9968000000000001, "central_musd": 4.4928, "high_musd": 9.984 }, { "component_id": "kimi-k2-base", "family": "Kimi K2 / K2.5", "organization": "Moonshot", "unit": "1e24_FLOP", "quantity": 2.976, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "32B active×15.5T.", "sources": [ "https://huggingface.co/moonshotai/Kimi-K2-Base", "https://arxiv.org/html/2602.02276" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 2.3808000000000002, "central_musd": 5.3568, "high_musd": 11.904 }, { "component_id": "kimi-k25-increment", "family": "Kimi K2 / K2.5", "organization": "Moonshot", "unit": "1e24_FLOP", "quantity": 2.88, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "32B active×15T ADDITIONAL tokens; base not counted again.", "sources": [ "https://huggingface.co/moonshotai/Kimi-K2-Base", "https://arxiv.org/html/2602.02276" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 2.304, "central_musd": 5.184, "high_musd": 11.52 }, { "component_id": "longcat-flash-20t-slice", "family": "LongCat Flash + 2.0", "organization": "Meituan", "unit": "1e24_FLOP", "quantity": 3.24, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "Average27B active×20T documented slice; no assumed extra3T.", "sources": [ "https://huggingface.co/meituan-longcat/LongCat-2.0", "https://arxiv.org/html/2509.01322" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 2.5920000000000005, "central_musd": 5.832000000000001, "high_musd": 12.96 }, { "component_id": "longcat2-35t-slice", "family": "LongCat Flash + 2.0", "organization": "Meituan", "unit": "1e24_FLOP", "quantity": 10.08, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "48B active×35T documented lower token slice; ASIC efficiency uncertainty not resolved.", "sources": [ "https://huggingface.co/meituan-longcat/LongCat-2.0", "https://arxiv.org/html/2509.01322" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 8.064, "central_musd": 18.144000000000002, "high_musd": 40.32 }, { "component_id": "ling1t-base", "family": "Ling 1T / 2.6", "organization": "Ant Group", "unit": "1e24_FLOP", "quantity": 6, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "50B active×20T.", "sources": [ "https://huggingface.co/inclusionAI/Ling-1T", "https://huggingface.co/inclusionAI/Ling-2.6-1T-base" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 4.800000000000001, "central_musd": 10.8, "high_musd": 24 }, { "component_id": "ling26-increment", "family": "Ling 1T / 2.6", "organization": "Ant Group", "unit": "1e24_FLOP", "quantity": 2.88, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "50B active×9.6T continuation.", "sources": [ "https://huggingface.co/inclusionAI/Ling-1T", "https://huggingface.co/inclusionAI/Ling-2.6-1T-base" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 2.304, "central_musd": 5.184, "high_musd": 11.52 }, { "component_id": "glm45-and-air", "family": "GLM 4.5 / Air", "organization": "Z.ai", "unit": "1e24_FLOP", "quantity": 6.0984, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "(32+12)B active×23.1T.", "sources": [ "https://huggingface.co/zai-org/GLM-4.5-Base" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 4.87872, "central_musd": 10.97712, "high_musd": 24.3936 }, { "component_id": "glm5-base", "family": "GLM 5 / 5.x", "organization": "Z.ai", "unit": "1e24_FLOP", "quantity": 6.84, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "40B active×28.5T; no later-version or RL margin.", "sources": [ "https://huggingface.co/zai-org/GLM-5" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 5.472, "central_musd": 12.312, "high_musd": 27.36 }, { "component_id": "hunyuan-large-reported-flops", "family": "Hunyuan Large / Hy3 / Hy4", "organization": "Tencent", "unit": "1e24_FLOP", "quantity": 3.49, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "Paper architecture-specific9.59ND; retain this rather than cheaper6ND proxy.", "sources": [ "https://arxiv.org/abs/2411.02265" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 2.7920000000000003, "central_musd": 6.282000000000001, "high_musd": 13.96 }, { "component_id": "nemotron3-ultra", "family": "Nemotron 3 / 3.5", "organization": "NVIDIA", "unit": "1e24_FLOP", "quantity": 6.6, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "55B active×20T; NVFP4 pricing not observed.", "sources": [ "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16", "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-Base-BF16", "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16", "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16", "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 5.28, "central_musd": 11.879999999999999, "high_musd": 26.4 }, { "component_id": "nemotron3-super", "family": "Nemotron 3 / 3.5", "organization": "NVIDIA", "unit": "1e24_FLOP", "quantity": 1.8, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "12B active×25T trained; unique-corpus size not used.", "sources": [ "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-Base-BF16" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 1.4400000000000002, "central_musd": 3.24, "high_musd": 7.2 }, { "component_id": "nemotron3-nano", "family": "Nemotron 3 / 3.5", "organization": "NVIDIA", "unit": "1e24_FLOP", "quantity": 0.48, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "3.2B active×25T trained, NOT10.6T unique corpus.", "sources": [ "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 0.384, "central_musd": 0.864, "high_musd": 1.92 }, { "component_id": "nemotron4-340b", "family": "Nemotron 4 / H", "organization": "NVIDIA", "unit": "1e24_FLOP", "quantity": 18.36, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "340B×9T(8T+1T).", "sources": [ "https://huggingface.co/nvidia/Nemotron-4-340B-Base", "https://arxiv.org/abs/2504.03624" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 14.688, "central_musd": 33.048, "high_musd": 73.44 }, { "component_id": "nemotron-h56b", "family": "Nemotron 4 / H", "organization": "NVIDIA", "unit": "1e24_FLOP", "quantity": 6.72, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "56B×20T; no pruned47B base duplicated.", "sources": [ "https://huggingface.co/nvidia/Nemotron-4-340B-Base", "https://arxiv.org/abs/2504.03624" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 5.376, "central_musd": 12.096, "high_musd": 26.88 }, { "component_id": "nemotron-h8b", "family": "Nemotron 4 / H", "organization": "NVIDIA", "unit": "1e24_FLOP", "quantity": 0.72, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "8B×15T; omitted in V1.", "sources": [ "https://huggingface.co/nvidia/Nemotron-4-340B-Base", "https://arxiv.org/abs/2504.03624" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 0.576, "central_musd": 1.296, "high_musd": 2.88 }, { "component_id": "minimax-text01", "family": "MiniMax Text-01 / M1", "organization": "MiniMax", "unit": "1e24_FLOP", "quantity": 3.13956, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "45.9B×11.4T(7.2+3.2+1).", "sources": [ "https://huggingface.co/MiniMaxAI/MiniMax-Text-01", "https://arxiv.org/html/2506.13585", "https://arxiv.org/html/2501.08313" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 2.511648, "central_musd": 5.651208, "high_musd": 12.55824 }, { "component_id": "minimax-m1-cpt", "family": "MiniMax Text-01 / M1", "organization": "MiniMax", "unit": "1e24_FLOP", "quantity": 2.0655, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "45.9B×7.5T incremental.", "sources": [ "https://huggingface.co/MiniMaxAI/MiniMax-Text-01", "https://arxiv.org/html/2506.13585", "https://arxiv.org/html/2501.08313" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 1.6524, "central_musd": 3.7179, "high_musd": 8.262 }, { "component_id": "minimax-m1-rl-published", "family": "MiniMax Text-01 / M1", "organization": "MiniMax", "unit": "million_USD_reported", "quantity": 0.53, "quantity_provenance": "Published rental-equivalent cost", "scope": "Separate reported RL estimate; not invoice. Reporting rate embedded.", "sources": [ "https://huggingface.co/MiniMaxAI/MiniMax-Text-01", "https://arxiv.org/html/2506.13585", "https://arxiv.org/html/2501.08313" ], "conversion_rates": [ 1, 1, 1 ], "low_musd": 0.53, "central_musd": 0.53, "high_musd": 0.53 }, { "component_id": "falcon-family", "family": "Falcon 7B / 40B / 180B", "organization": "TII", "unit": "1e24_FLOP", "quantity": 4.0614, "quantity_provenance": "180B reportedPF-days; smaller6ND", "scope": "180B3.7584e24 +40B.24 +7B.063. Older-generation conversion prices.", "sources": [ "https://arxiv.org/html/2311.16867", "https://huggingface.co/tiiuae/falcon-180B" ], "conversion_rates": [ 1.5, 3, 6 ], "low_musd": 6.0921, "central_musd": 12.1842, "high_musd": 24.3684 }, { "component_id": "gemma2-three-sizes", "family": "Gemma 2", "organization": "Google", "unit": "1e24_FLOP", "quantity": 2.562, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "27B13T +9B8T +2B2T; teacher forward cost unpriced.", "sources": [ "https://arxiv.org/html/2408.00118" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 2.0496, "central_musd": 4.6116, "high_musd": 10.248 }, { "component_id": "gemma3-four-sizes", "family": "Gemma 3", "organization": "Google", "unit": "1e24_FLOP", "quantity": 3.24, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "27B14T +12B12T +4B4T +1B2T; no3n/270M extrapolation.", "sources": [ "https://arxiv.org/html/2503.19786" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 2.5920000000000005, "central_musd": 5.832000000000001, "high_musd": 12.96 }, { "component_id": "gptoss-two-sizes", "family": "gpt-oss 120B / 20B", "organization": "OpenAI", "unit": "million_H100_hours", "quantity": 2.31, "quantity_provenance": "Published hardware-hours; rental price is assumed", "scope": "120B2.1M;20B almost10×less, approximated. No extra posttraining margin.", "sources": [ "https://arxiv.org/html/2508.10925" ], "conversion_rates": [ 1.5, 2.5, 4 ], "low_musd": 3.465, "central_musd": 5.775, "high_musd": 9.24 }, { "component_id": "olmo2-three-sizes", "family": "OLMo 2", "organization": "Ai2", "unit": "1e24_FLOP", "quantity": 1.71, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "7B4T+13B5T+32B6T.", "sources": [ "https://huggingface.co/allenai/OLMo-2-0325-32B" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 1.368, "central_musd": 3.078, "high_musd": 6.84 }, { "component_id": "granite3-dense", "family": "Granite 3 dense", "organization": "IBM", "unit": "1e24_FLOP", "quantity": 0.72, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "8B+2B,12T each; other generations not included.", "sources": [ "https://huggingface.co/ibm-granite/granite-3.0-8b-base" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 0.576, "central_musd": 1.296, "high_musd": 2.88 }, { "component_id": "phi4-14b", "family": "Phi 4", "organization": "Microsoft", "unit": "1e24_FLOP", "quantity": 0.8232, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "14B9.8T; Mini not included.", "sources": [ "https://huggingface.co/microsoft/phi-4" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 0.65856, "central_musd": 1.4817600000000002, "high_musd": 3.2928 }, { "component_id": "trinity-large-final", "family": "Trinity Large", "organization": "Arcee / Prime Intellect", "unit": "1e24_FLOP", "quantity": 1.326, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "13B active×17T final;10T checkpoint not added.", "sources": [ "https://arxiv.org/html/2602.17004", "https://huggingface.co/arcee-ai/Trinity-Large-TrueBase" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 1.0608000000000002, "central_musd": 2.3868, "high_musd": 5.304 }, { "component_id": "inkling-large", "family": "Inkling / Small", "organization": "Thinking Machines", "unit": "1e24_FLOP", "quantity": 11.07, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "41B active×45T from launch post; Small excluded.", "sources": [ "https://thinkingmachines.ai/news/introducing-inkling/" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 8.856, "central_musd": 19.926000000000002, "high_musd": 44.28 }, { "component_id": "mimo-three-bases", "family": "MiMo 7B / V2 Flash / V2.5 Pro", "organization": "Xiaomi", "unit": "1e24_FLOP", "quantity": 10.284, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "7B25T +15B27T +42B27T.", "sources": [ "https://huggingface.co/XiaomiMiMo/MiMo-V2-Flash", "https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro", "https://arxiv.org/abs/2505.07608" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 8.227200000000002, "central_musd": 18.511200000000002, "high_musd": 41.136 }, { "component_id": "telechat2-115b", "family": "TeleChat 1 / 2", "organization": "China Telecom", "unit": "1e24_FLOP", "quantity": 6.9, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "115B10T primary card; smaller models omitted from documented slice.", "sources": [ "https://huggingface.co/Tele-AI/TeleChat2-115B", "https://github.com/Tele-AI/TeleChat2" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 5.5200000000000005, "central_musd": 12.420000000000002, "high_musd": 27.6 }, { "component_id": "apertus70-6nd", "family": "Apertus 8B / 70B", "organization": "Swiss AI", "unit": "1e24_FLOP", "quantity": 6.3, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "70B15T primary card;8B and reported6.74e24 secondary value not assumed here.", "sources": [ "https://huggingface.co/swiss-ai/Apertus-70B-2509", "https://github.com/swiss-ai/apertus-tech-report/blob/main/Apertus_Tech_Report.pdf" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 5.04, "central_musd": 11.34, "high_musd": 25.2 }, { "component_id": "solar-open2-250b", "family": "Solar Open 100B / Open2 250B", "organization": "Upstage", "unit": "million_B200_hours", "quantity": 2, "quantity_provenance": "Published hardware-hours; rental price is assumed", "scope": "Primary card2M GPU-hours; divergent Epoch$32M not added.", "sources": [ "https://huggingface.co/upstage/Solar-Open2-250B", "https://huggingface.co/upstage/Solar-Open-100B", "https://epoch.ai/data/ai-models" ], "conversion_rates": [ 3, 5, 8 ], "low_musd": 6, "central_musd": 10, "high_musd": 16 }, { "component_id": "solar-open100b", "family": "Solar Open 100B / Open2 250B", "organization": "Upstage", "unit": "1e24_FLOP", "quantity": 1.4184, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "12B active×19.7T.", "sources": [ "https://huggingface.co/upstage/Solar-Open2-250B", "https://huggingface.co/upstage/Solar-Open-100B", "https://epoch.ai/data/ai-models" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 1.1347200000000002, "central_musd": 2.5531200000000003, "high_musd": 5.6736 }, { "component_id": "dbrx-base", "family": "DBRX", "organization": "Databricks", "unit": "1e24_FLOP", "quantity": 2.592, "quantity_provenance": "Reported parameters/tokens;6ND proxy unless specified", "scope": "36B active×12T.", "sources": [ "https://www.databricks.com/blog/introducing-dbrx-new-state-art-open-llm", "https://huggingface.co/databricks/dbrx-base" ], "conversion_rates": [ 0.8, 1.8, 4 ], "low_musd": 2.0736000000000003, "central_musd": 4.6656, "high_musd": 10.368 } ]