YFarmXIntelligence for an accelerating worldBriefings
YFarmX
YFarmX
AI Desk · Model Identity

Identity watch

A model can change behind its name. The watch records what appeared, what measured differently, and what measurement placed, so a silent change stops being silent.

What changed

Every row is a measured difference between two dated snapshots: the value it was, and the value it became

3 September 2026

AionLabs: Aion-RP 1.0 (8B)aion-labs/aion-rp-llama-3.1-8bMax output: 32,768 became 29,491OutputAllenAI: Olmo 3 32B Thinkallenai/olmo-3-32b-thinkLeft the catalogue: none became noneRemovedAnthropic: Claude Fable Latest~anthropic/claude-fable-latestSupported parameters: include_reasoning, max_completion_tokens, max_tokens, reasoning, reasoning_effort, response_format, stop, structured_outputs, tool_choice, tools, verbosity became include_reasoning, max_completion_tokens, max_tokens, reasoning, reasoning_effort, response_format, stop, structured_outputs, tools, verbosityAPIAnthropic: Claude Opus 4.7 (Fast)anthropic/claude-opus-4.7-fastLeft the catalogue: none became noneRemovedAnthropic: Claude Opus 4.8 (Fast)anthropic/claude-opus-4.8-fastLeft the catalogue: none became noneRemovedArcee AI: Trinity Large Thinkingarcee-ai/trinity-large-thinkingMax output: 262,144 became 80,000OutputArcee AI: Trinity Large Thinkingarcee-ai/trinity-large-thinkingPrompt price / token: $0.22 / M tokens became $0.25 / M tokensPriceArcee AI: Trinity Large Thinkingarcee-ai/trinity-large-thinkingCompletion price / token: $0.85 / M tokens became $0.80 / M tokensPriceArcee AI: Trinity Large Thinkingarcee-ai/trinity-large-thinkingSupported parameters: frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p became include_reasoning, max_tokens, reasoning, temperature, tool_choice, tools, top_k, top_pAPIArcee AI: Virtuoso Largearcee-ai/virtuoso-largeLeft the catalogue: none became noneRemovedByteDance Seed: Seed 2.1 Turbobytedance-seed/seed-2-1-turboMax output: 262,144 became 235,929OutputClaude Opus 5 (Fast)anthropic/claude-opus-5-fastLeft the catalogue: none became noneRemovedDeep Cogito: Cogito v2.1 671Bdeepcogito/cogito-v2.1-671bLeft the catalogue: none became noneRemovedDeepSeek V4 Flash Latest~deepseek/deepseek-v4-flash-latestMax output: 1,048,576 became 393,216OutputDeepSeek V4 Flash Latest~deepseek/deepseek-v4-flash-latestPrompt price / token: $0.07 / M tokens became $0.05 / M tokensPriceDeepSeek V4 Flash Latest~deepseek/deepseek-v4-flash-latestCompletion price / token: $0.18 / M tokens became $0.16 / M tokensPriceDeepSeek: DeepSeek V3deepseek/deepseek-chatMax output: 16,000 became 16,384OutputDeepSeek: DeepSeek V3deepseek/deepseek-chatPrompt price / token: $0.26 / M tokens became $0.32 / M tokensPriceDeepSeek: DeepSeek V3deepseek/deepseek-chatCompletion price / token: $1.03 / M tokens became $0.89 / M tokensPriceDeepSeek: DeepSeek V3 0324deepseek/deepseek-chat-v3-0324Max output: 163,840 became 147,456OutputDeepSeek: DeepSeek V3.1deepseek/deepseek-chat-v3.1Max output: 32,768 became 144,900OutputDeepSeek: DeepSeek V3.1deepseek/deepseek-chat-v3.1Prompt price / token: $0.25 / M tokens became $0.55 / M tokensPriceDeepSeek: DeepSeek V3.1deepseek/deepseek-chat-v3.1Completion price / token: $0.95 / M tokens became $1.65 / M tokensPriceDeepSeek: DeepSeek V4 Flash 0423deepseek/deepseek-v4-flashPrompt price / token: $0.08 / M tokens became $0.08 / M tokensPriceDeepSeek: DeepSeek V4 Flash 0423deepseek/deepseek-v4-flashCompletion price / token: $0.16 / M tokens became $0.16 / M tokensPriceDeepSeek: DeepSeek V4 Flash 0731deepseek/deepseek-v4-flash-0731Max output: 384,000 became 943,718OutputDeepSeek: DeepSeek V4 Flash 0731deepseek/deepseek-v4-flash-0731Prompt price / token: $0.08 / M tokens became $0.07 / M tokensPriceDeepSeek: DeepSeek V4 Flash Vision Expdeepseek/deepseek-v4-flash-vision-expSupported parameters: frequency_penalty, include_reasoning, logprobs, max_tokens, presence_penalty, reasoning, reasoning_effort, response_format, stop, temperature, tool_choice, tools, top_logprobs, top_p became frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, reasoning_effort, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_pAPIDeepSeek: DeepSeek V4 Pro 0423deepseek/deepseek-v4-proMax output: 393,216 became 384,000OutputDeepSeek: DeepSeek V4 Pro 0423deepseek/deepseek-v4-proPrompt price / token: $1.60 / M tokens became $1.03 / M tokensPriceDeepSeek: DeepSeek V4 Pro 0423deepseek/deepseek-v4-proCompletion price / token: $3.20 / M tokens became $2.06 / M tokensPriceDeepSeek: DeepSeek V4 Pro 0813deepseek/deepseek-v4-pro-0813Max output: none became 384,000OutputDeepSeek: DeepSeek V4 Pro 0813deepseek/deepseek-v4-pro-0813Prompt price / token: $1.19 / M tokens became $0.66 / M tokensPriceDeepSeek: DeepSeek V4 Pro 0813deepseek/deepseek-v4-pro-0813Completion price / token: $3.56 / M tokens became $1.98 / M tokensPriceDeepSeek: R1 Distill Llama 70Bdeepseek/deepseek-r1-distill-llama-70bMax output: 8,192 became 7,372OutputDots Studio: Dots3-Note Preview (free)dots-studio/dots-3-note-preview:freeMax output: 512,000 became 460,800OutputGoogle Gemini Flash Latest~google/gemini-flash-latestPrompt price / token: $0.38 / M tokens became $0.75 / M tokensPriceGoogle Gemini Flash Latest~google/gemini-flash-latestCompletion price / token: $1.88 / M tokens became $3.75 / M tokensPriceGoogle: Gemini 3.7 Flashgoogle/gemini-3.7-flashPrompt price / token: $0.38 / M tokens became $0.75 / M tokensPriceGoogle: Gemini 3.7 Flashgoogle/gemini-3.7-flashCompletion price / token: $1.88 / M tokens became $3.75 / M tokensPriceGoogle: Gemini 3.7 Flash (batch)google/gemini-3.7-flash:batchPrompt price / token: $0.19 / M tokens became $0.38 / M tokensPriceGoogle: Gemini 3.7 Flash (batch)google/gemini-3.7-flash:batchCompletion price / token: $0.94 / M tokens became $1.88 / M tokensPriceGoogle: Gemma 3 27Bgoogle/gemma-3-27b-itContext window: 262,144 became 131,072ContextGoogle: Gemma 3 27Bgoogle/gemma-3-27b-itMax output: 131,072 became 117,964OutputGoogle: Gemma 3n 4Bgoogle/gemma-3n-e4b-itLeft the catalogue: none became noneRemovedGoogle: Gemma 4 31Bgoogle/gemma-4-31b-itMax output: 262,144 became 16,384OutputGoogle: Gemma 4 31Bgoogle/gemma-4-31b-itPrompt price / token: $0.10 / M tokens became $0.09 / M tokensPriceGoogle: Nano Banana 2 (Gemini 3.1 Flash Image Preview)google/gemini-3.1-flash-image-previewMax output: 65,536 became 58,982OutputGoogle: Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)google/gemini-3.1-flash-lite-imageMax output: 65,536 became 58,982OutputIBM: Granite 4.0 Microibm-granite/granite-4.0-h-microMax output: 131,000 became 117,900OutputIBM: Granite 4.1 8Bibm-granite/granite-4.1-8bMax output: 131,072 became 117,964OutputinclusionAI: Ling-2.6-1Tinclusionai/ling-2.6-1tLeft the catalogue: none became noneRemovedinclusionAI: Ling-2.6-flashinclusionai/ling-2.6-flashLeft the catalogue: none became noneRemovedinclusionAI: Ring-2.6-1Tinclusionai/ring-2.6-1tLeft the catalogue: none became noneRemovedKwaipilot: KAT-Coder-Air V2.5kwaipilot/kat-coder-air-v2.5Left the catalogue: none became noneRemovedKwaipilot: KAT-Coder-Pro V2kwaipilot/kat-coder-pro-v2Max output: 80,000 became 144,000OutputKwaipilot: KAT-Coder-Pro V2kwaipilot/kat-coder-pro-v2Supported parameters: frequency_penalty, logit_bias, logprobs, max_tokens, min_p, presence_penalty, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p became frequency_penalty, logit_bias, max_tokens, min_p, presence_penalty, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_pAPIKwaipilot: KAT-Coder-Pro V2.5kwaipilot/kat-coder-pro-v2.5Context window: 256,000 became 262,144ContextKwaipilot: KAT-Coder-Pro V2.5kwaipilot/kat-coder-pro-v2.5Max output: 80,000 became 235,929OutputKwaipilot: KAT-Coder-Pro V2.5kwaipilot/kat-coder-pro-v2.5Supported parameters: logprobs, max_tokens, presence_penalty, response_format, stop, structured_outputs, temperature, tool_choice, tools, top_logprobs, top_p became frequency_penalty, logit_bias, max_tokens, min_p, presence_penalty, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_pAPIMagnum v4 72Banthracite-org/magnum-v4-72bPrompt price / token: $3.00 / M tokens became $2.50 / M tokensPriceMancer: Weaver (alpha)mancer/weaverPrompt price / token: $0.50 / M tokens became $0.40 / M tokensPriceMeta: Llama 3.1 8B Instructmeta-llama/llama-3.1-8b-instructMax output: 131,072 became 117,964OutputMeta: Llama 3.2 1B Instructmeta-llama/llama-3.2-1b-instructMax output: 60,000 became 54,000OutputMeta: Llama 3.2 3B Instructmeta-llama/llama-3.2-3b-instructMax output: 131,072 became 117,964OutputMeta: Llama 4 Maverickmeta-llama/llama-4-maverickMax output: 16,384 became 115,200OutputMeta: Llama 4 Maverickmeta-llama/llama-4-maverickCompletion price / token: $0.80 / M tokens became $0.70 / M tokensPriceMeta: Llama Guard 4 12Bmeta-llama/llama-guard-4-12bContext window: 1,048,576 became 163,840ContextMeta: Muse Glimmer 30Bmeta/muse-glimmer-30bMax output: 131,072 became 117,964OutputMeta: Muse Spark 1.1meta/muse-spark-1.1Max output: none became 943,718OutputMeta: Muse Spark 1.2meta/muse-spark-1.2Max output: none became 943,718OutputMicrosoft: Phi 4microsoft/phi-4Max output: 16,384 became 14,745OutputMiniMax: MiniMax M2.5minimax/minimax-m2.5Max output: 32,768 became 128,000OutputMiniMax: MiniMax M2.5minimax/minimax-m2.5Completion price / token: $0.95 / M tokens became $1.08 / M tokensPriceMiniMax: MiniMax M3 (batch)minimax/minimax-m3:batchMax output: none became 471,859OutputMiniMax: MiniMax-01minimax/minimax-01Max output: 1,000,192 became 900,172OutputMistral Largemistralai/mistral-largeMax output: none became 102,400OutputMistral Large 2407mistralai/mistral-large-2407Max output: none became 104,857OutputMistral: Codestral 2508mistralai/codestral-2508Max output: none became 204,800OutputMistral: Ministral 3 14B 2512mistralai/ministral-14b-2512Max output: none became 209,715OutputMistral: Ministral 3 3B 2512mistralai/ministral-3b-2512Max output: none became 104,857OutputMistral: Ministral 3 8B 2512mistralai/ministral-8b-2512Max output: none became 209,715OutputMistral: Ministral 8Bmistralai/ministral-8bLeft the catalogue: none became noneRemovedMistral: Mistral Large 3 2512mistralai/mistral-large-2512Max output: none became 209,715OutputMistral: Mistral Medium 3mistralai/mistral-medium-3Max output: none became 104,857OutputMistral: Mistral Medium 3.1mistralai/mistral-medium-3.1Max output: none became 104,857OutputMistral: Mistral Medium 3.5mistralai/mistral-medium-3-5Max output: none became 209,715OutputMistral: Mistral Small 3.1 24Bmistralai/mistral-small-3.1-24b-instructMax output: 128,000 became 102,400OutputMistral: Mistral Small 3.2 24Bmistralai/mistral-small-3.2-24b-instructContext window: 256,000 became 131,072ContextMistral: Mistral Small 3.2 24Bmistralai/mistral-small-3.2-24b-instructPrompt price / token: $0.09 / M tokens became $0.07 / M tokensPriceMistral: Mistral Small 3.2 24Bmistralai/mistral-small-3.2-24b-instructCompletion price / token: $0.25 / M tokens became $0.20 / M tokensPriceMistral: Mistral Small 4mistralai/mistral-small-2603Max output: none became 209,715OutputMistral: Mixtral 8x22B Instructmistralai/mixtral-8x22b-instructMax output: none became 52,428OutputMistral: Sabamistralai/mistral-sabaMax output: none became 26,214OutputMistral: Voxtral Small 24B 2507mistralai/voxtral-small-24b-2507Context window: 32,000 became 32,768ContextMistral: Voxtral Small 24B 2507mistralai/voxtral-small-24b-2507Max output: none became 26,214OutputMoonshotAI Kimi Latest~moonshotai/kimi-latestMax output: 974,842 became 943,718OutputMoonshotAI Kimi Latest~moonshotai/kimi-latestPrompt price / token: $2.60 / M tokens became $2.50 / M tokensPriceMoonshotAI Kimi Latest~moonshotai/kimi-latestCompletion price / token: $13.0 / M tokens became $14.0 / M tokensPriceMoonshotAI: Kimi K2.5moonshotai/kimi-k2.5Max output: 262,144 became 235,929OutputMoonshotAI: Kimi K2.6moonshotai/kimi-k2.6Max output: 262,144 became 235,929OutputMoonshotAI: Kimi K2.6moonshotai/kimi-k2.6Prompt price / token: $0.58 / M tokens became $0.95 / M tokensPriceMoonshotAI: Kimi K2.6moonshotai/kimi-k2.6Completion price / token: $2.44 / M tokens became $4.00 / M tokensPriceMoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-codeMax output: 262,144 became 235,929OutputMoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-codePrompt price / token: $0.67 / M tokens became $0.66 / M tokensPriceMoonshotAI: Kimi K2.7 Code (batch)moonshotai/kimi-k2.7-code:batchLeft the catalogue: none became noneRemovedMoonshotAI: Kimi K3moonshotai/kimi-k3Max output: none became 943,718OutputMythoMax 13Bgryphe/mythomax-l2-13bMax output: 4,096 became 3,686OutputNex AGI: Nex-N2-Mininex-agi/nex-n2-miniMax output: 262,144 became 235,929OutputNex AGI: Nex-N2-Pronex-agi/nex-n2-proMax output: 262,144 became 235,929OutputNous: Hermes 4 405Bnousresearch/hermes-4-405bMax output: none became 117,964OutputNous: Hermes 4 70Bnousresearch/hermes-4-70bMax output: none became 117,964OutputNVIDIA: Nemotron 3 Nano 30B A3Bnvidia/nemotron-3-nano-30b-a3bMax output: 262,144 became 235,929OutputNVIDIA: Nemotron 3 Nano 30B A3B (free)nvidia/nemotron-3-nano-30b-a3b:freeLeft the catalogue: none became noneRemovedNVIDIA: Nemotron 3 Super (free)nvidia/nemotron-3-super-120b-a12b:freeMax output: 262,144 became 235,929OutputNVIDIA: Nemotron 3 Ultranvidia/nemotron-3-ultra-550b-a55bContext window: 512,288 became 262,144ContextNVIDIA: Nemotron 3 Ultranvidia/nemotron-3-ultra-550b-a55bMax output: none became 182,520OutputNVIDIA: Nemotron 3 Ultranvidia/nemotron-3-ultra-550b-a55bCompletion price / token: $3.60 / M tokens became $2.40 / M tokensPriceNVIDIA: Nemotron 3 Ultra (batch)nvidia/nemotron-3-ultra-550b-a55b:batchLeft the catalogue: none became noneRemovedNVIDIA: Nemotron Nano 12B 2 VL (free)nvidia/nemotron-nano-12b-v2-vl:freeLeft the catalogue: none became noneRemovedNVIDIA: Nemotron Nano 9B V2 (free)nvidia/nemotron-nano-9b-v2:freeLeft the catalogue: none became noneRemovedOpenAI GPT Latest~openai/gpt-latestPrompt price / token: $2.50 / M tokens became $2.00 / M tokensPriceOpenAI GPT Latest~openai/gpt-latestCompletion price / token: $15.0 / M tokens became $10.0 / M tokensPriceOpenAI: GPT-3.5 Turbo (older v0613)openai/gpt-3.5-turbo-0613Max output: 4,096 became 3,685OutputOpenAI: GPT-3.5 Turbo Instructopenai/gpt-3.5-turbo-instructMax output: 4,096 became 3,685OutputOpenAI: GPT-5 Codex (batch)openai/gpt-5-codex:batchLeft the catalogue: none became noneRemovedOpenAI: GPT-5.6 Solopenai/gpt-5.6-solPrompt price / token: $2.50 / M tokens became $2.00 / M tokensPriceOpenAI: GPT-5.6 Solopenai/gpt-5.6-solCompletion price / token: $15.0 / M tokens became $10.0 / M tokensPriceOpenAI: GPT-5.6 Sol (batch)openai/gpt-5.6-sol:batchPrompt price / token: $1.25 / M tokens became $1.00 / M tokensPriceOpenAI: GPT-5.6 Sol (batch)openai/gpt-5.6-sol:batchCompletion price / token: $7.50 / M tokens became $5.00 / M tokensPriceOpenAI: GPT-5.6 Sol Proopenai/gpt-5.6-sol-proPrompt price / token: $2.50 / M tokens became $2.00 / M tokensPriceOpenAI: GPT-5.6 Sol Proopenai/gpt-5.6-sol-proCompletion price / token: $15.0 / M tokens became $10.0 / M tokensPriceOpenAI: GPT-5.6 Sol Pro (batch)openai/gpt-5.6-sol-pro:batchPrompt price / token: $1.25 / M tokens became $1.00 / M tokensPriceOpenAI: GPT-5.6 Sol Pro (batch)openai/gpt-5.6-sol-pro:batchCompletion price / token: $7.50 / M tokens became $5.00 / M tokensPriceOpenAI: gpt-oss-120bopenai/gpt-oss-120bMax output: 131,072 became 117,964OutputOpenAI: gpt-oss-120bopenai/gpt-oss-120bPrompt price / token: $0.03 / M tokens became $0.04 / M tokensPriceOpenAI: gpt-oss-20bopenai/gpt-oss-20bMax output: 131,072 became 117,964OutputOpenAI: gpt-oss-20b (free)openai/gpt-oss-20b:freeLeft the catalogue: none became noneRemovedOpenAI: o1 (batch)openai/o1:batchLeft the catalogue: none became noneRemovedOpenAI: o1-pro (batch)openai/o1-pro:batchLeft the catalogue: none became noneRemovedOpenAI: o3 Mini High (batch)openai/o3-mini-high:batchLeft the catalogue: none became noneRemovedOpenAI: o3 Pro (batch)openai/o3-pro:batchLeft the catalogue: none became noneRemovedOpenAI: o4 Mini High (batch)openai/o4-mini-high:batchLeft the catalogue: none became noneRemovedOx Alphastealth/ox-alphaLeft the catalogue: none became noneRemovedPerplexity: Sonarperplexity/sonarMax output: none became 114,364OutputPerplexity: Sonar Deep Researchperplexity/sonar-deep-researchMax output: none became 115,200OutputPerplexity: Sonar Reasoning Properplexity/sonar-reasoning-proMax output: none became 115,200OutputQwen: Qwen Plus 0728 (thinking)qwen/qwen-plus-2025-07-28:thinkingLeft the catalogue: none became noneRemovedQwen: Qwen2.5 7B Instructqwen/qwen-2.5-7b-instructMax output: 32,768 became 29,491OutputQwen: Qwen2.5 7B Instructqwen/qwen-2.5-7b-instructSupported parameters: frequency_penalty, logit_bias, max_tokens, min_p, presence_penalty, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_p became frequency_penalty, max_tokens, min_p, presence_penalty, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_pAPIQwen: Qwen2.5 VL 72B Instructqwen/qwen2.5-vl-72b-instructMax output: 128,000 became 115,200OutputQwen: Qwen3 235B A22B Instruct 2507qwen/qwen3-235b-a22b-2507Max output: 16,384 became 235,929OutputQwen: Qwen3 235B A22B Instruct 2507qwen/qwen3-235b-a22b-2507Prompt price / token: $0.09 / M tokens became $0.09 / M tokensPriceQwen: Qwen3 235B A22B Instruct 2507qwen/qwen3-235b-a22b-2507Completion price / token: $0.55 / M tokens became $0.35 / M tokensPriceQwen: Qwen3 235B A22B Thinking 2507qwen/qwen3-235b-a22b-thinking-2507Context window: 262,144 became 131,072ContextQwen: Qwen3 235B A22B Thinking 2507qwen/qwen3-235b-a22b-thinking-2507Max output: none became 117,964OutputQwen: Qwen3 235B A22B Thinking 2507qwen/qwen3-235b-a22b-thinking-2507Supported parameters: frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, temperature, tool_choice, tools, top_k, top_logprobs, top_p became frequency_penalty, include_reasoning, logprobs, max_tokens, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, temperature, tool_choice, tools, top_k, top_logprobs, top_pAPIQwen: Qwen3 30B A3Bqwen/qwen3-30b-a3bMax output: 8,192 became 16,384OutputQwen: Qwen3 30B A3Bqwen/qwen3-30b-a3bPrompt price / token: $0.13 / M tokens became $0.12 / M tokensPriceQwen: Qwen3 30B A3Bqwen/qwen3-30b-a3bCompletion price / token: $0.52 / M tokens became $0.50 / M tokensPriceQwen: Qwen3 32Bqwen/qwen3-32bSupported parameters: frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p became frequency_penalty, include_reasoning, logit_bias, max_tokens, min_p, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_pAPIQwen: Qwen3 Coder 30B A3B Instructqwen/qwen3-coder-30b-a3b-instructMax output: 262,144 became 235,929OutputQwen: Qwen3 Coder Nextqwen/qwen3-coder-nextMax output: 262,144 became 235,929OutputQwen: Qwen3 Next 80B A3B Instructqwen/qwen3-next-80b-a3b-instructMax output: 16,384 became 235,929OutputQwen: Qwen3 Next 80B A3B Instructqwen/qwen3-next-80b-a3b-instructPrompt price / token: $0.09 / M tokens became $0.10 / M tokensPriceQwen: Qwen3 Next 80B A3B Thinkingqwen/qwen3-next-80b-a3b-thinkingSupported parameters: frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p became frequency_penalty, include_reasoning, logprobs, max_tokens, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_pAPIQwen: Qwen3 VL 30B A3B Instructqwen/qwen3-vl-30b-a3b-instructMax output: 32,768 became 16,384OutputQwen: Qwen3 VL 30B A3B Instructqwen/qwen3-vl-30b-a3b-instructPrompt price / token: $0.13 / M tokens became $0.15 / M tokensPriceQwen: Qwen3 VL 30B A3B Instructqwen/qwen3-vl-30b-a3b-instructCompletion price / token: $0.52 / M tokens became $0.60 / M tokensPriceQwen: Qwen3.5 397B A17Bqwen/qwen3.5-397b-a17bMax output: 65,536 became 235,929OutputQwen: Qwen3.5 397B A17Bqwen/qwen3.5-397b-a17bPrompt price / token: $0.39 / M tokens became $0.55 / M tokensPriceQwen: Qwen3.5 397B A17Bqwen/qwen3.5-397b-a17bCompletion price / token: $2.34 / M tokens became $3.50 / M tokensPriceQwen: Qwen3.5-122B-A10Bqwen/qwen3.5-122b-a10bMax output: 262,144 became 81,920OutputQwen: Qwen3.5-122B-A10Bqwen/qwen3.5-122b-a10bPrompt price / token: $0.26 / M tokens became $0.29 / M tokensPriceQwen: Qwen3.5-122B-A10Bqwen/qwen3.5-122b-a10bCompletion price / token: $2.08 / M tokens became $2.40 / M tokensPriceQwen: Qwen3.5-35B-A3Bqwen/qwen3.5-35b-a3bMax output: 262,144 became 235,929OutputQwen: Qwen3.5-9Bqwen/qwen3.5-9bMax output: 262,144 became 235,929OutputQwen: Qwen3.6 27Bqwen/qwen3.6-27bMax output: 262,144 became 235,929OutputQwen: Qwen3.6 35B A3Bqwen/qwen3.6-35b-a3bMax output: 262,144 became 235,929OutputQwen: Qwen3.6 35B A3Bqwen/qwen3.6-35b-a3bPrompt price / token: $0.14 / M tokens became $0.10 / M tokensPriceQwen: Qwen3.6 35B A3Bqwen/qwen3.6-35b-a3bCompletion price / token: $1.00 / M tokens became $0.90 / M tokensPriceQwen: Qwen3.8 2.4T A95Bqwen/qwen3.8-2.4t-a95bMax output: 131,072 became 262,144OutputQwen: Qwen3.8 27Bqwen/qwen3.8-27bPrompt price / token: $0.45 / M tokens became $0.42 / M tokensPriceQwen: Qwen3.8 27Bqwen/qwen3.8-27bCompletion price / token: $3.20 / M tokens became $3.00 / M tokensPriceQwen: Qwen3.8 27Bqwen/qwen3.8-27bSupported parameters: frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, presence_penalty, reasoning, reasoning_effort, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p became frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, reasoning_effort, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_pAPIQwen2.5 Coder 32B Instructqwen/qwen-2.5-coder-32b-instructMax output: 32,768 became 29,491OutputReka Edgerekaai/reka-edgeMax output: 16,384 became 14,745OutputReka Flash 3rekaai/reka-flash-3Max output: 65,536 became 58,982OutputReMM SLERP 13Bundi95/remm-slerp-l2-13bMax output: 6,144 became 4,096OutputSao10K: Llama 3 8B Lunarissao10k/l3-lunaris-8bMax output: 16,384 became 7,372OutputSpaceXAI: Grok 4.20x-ai/grok-4.20Max output: none became 1,800,000OutputSpaceXAI: Grok 4.20 Multi-Agentx-ai/grok-4.20-multi-agentMax output: none became 1,800,000OutputSpaceXAI: Grok 4.3x-ai/grok-4.3Max output: none became 900,000OutputSpaceXAI: Grok 4.5x-ai/grok-4.5Max output: none became 450,000OutputSpaceXAI: Grok 4.6x-ai/grok-4.6Max output: none became 450,000OutputSpaceXAI: Grok 4.6x-ai/grok-4.6Supported parameters: include_reasoning, logprobs, max_tokens, reasoning, reasoning_effort, response_format, seed, structured_outputs, temperature, tool_choice, tools, top_logprobs, top_p became include_reasoning, logprobs, max_tokens, reasoning, reasoning_effort, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_pAPISpaceXAI: Grok Build 0.1x-ai/grok-build-0.1Max output: none became 230,400OutputStepFun: Step 3.7 Flashstepfun/step-3.7-flashMax output: 256,000 became 230,400OutputTencent: Hunyuan A13B Instructtencent/hunyuan-a13b-instructMax output: 131,072 became 117,964OutputTencent: Hy3tencent/hy3Prompt price / token: $0.13 / M tokens became $0.08 / M tokensPriceTencent: Hy3tencent/hy3Completion price / token: $0.53 / M tokens became $0.33 / M tokensPriceTencent: Hy3 previewtencent/hy3-previewMax output: none became 235,929OutputTheDrummer: Cydonia 24B V4.1thedrummer/cydonia-24b-v4.1Max output: 131,072 became 117,964OutputTheDrummer: Rocinante 12Bthedrummer/rocinante-12bLeft the catalogue: none became noneRemovedTheDrummer: Skyfall 36B V2thedrummer/skyfall-36b-v2Max output: 32,768 became 29,491OutputTheDrummer: UnslopNemo 12Bthedrummer/unslopnemo-12bMax output: 1,024,000 became 26,214OutputThinking Machines: Inklingthinkingmachines/inklingMax output: 262,144 became 471,859OutputThinking Machines: Inklingthinkingmachines/inklingPrompt price / token: $0.95 / M tokens became $1.00 / M tokensPriceThinking Machines: Inkling (batch)thinkingmachines/inkling:batchMax output: none became 471,859OutputThinking Machines: Inkling (free)thinkingmachines/inkling:freeContext window: 262,144 became 1,048,576ContextThinking Machines: Inkling Smallthinkingmachines/inkling-smallSupported parameters: frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, reasoning_effort, repetition_penalty, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p became frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, reasoning_effort, repetition_penalty, seed, stop, temperature, tool_choice, tools, top_k, top_logprobs, top_pAPIThinking Machines: Inkling Small (free)thinkingmachines/inkling-small:freeContext window: 262,144 became 1,048,576ContextUpstage: Solar Pro 3upstage/solar-pro-3Max output: 131,072 became 117,964OutputxAI: Grok Latest~x-ai/grok-latestMax output: none became 450,000OutputxAI: Grok Latest~x-ai/grok-latestSupported parameters: include_reasoning, logprobs, max_tokens, reasoning, reasoning_effort, response_format, seed, structured_outputs, temperature, tool_choice, tools, top_logprobs, top_p became include_reasoning, logprobs, max_tokens, reasoning, reasoning_effort, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_pAPIXiaomi: MiMo-V2.5xiaomi/mimo-v2.5Supported parameters: frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p became frequency_penalty, include_reasoning, logit_bias, max_tokens, min_p, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_pAPIZ.ai: GLM 4.6z-ai/glm-4.6Prompt price / token: $0.50 / M tokens became $0.55 / M tokensPriceZ.ai: GLM 4.6z-ai/glm-4.6Completion price / token: $2.00 / M tokens became $2.20 / M tokensPriceZ.ai: GLM 4.7z-ai/glm-4.7Supported parameters: frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_a, top_k, top_logprobs, top_p became frequency_penalty, include_reasoning, logit_bias, max_tokens, min_p, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_a, top_k, top_pAPIZ.ai: GLM 5.2 (batch)z-ai/glm-5.2:batchLeft the catalogue: none became noneRemovedZ.ai: GLM 5.2 (free)z-ai/glm-5.2:freeMax output: 256,000 became 230,400OutputZ.ai: GLM 5.3z-ai/glm-5.3Context window: 1,048,576 became 1,310,720ContextZ.ai: GLM 5.3z-ai/glm-5.3Max output: 131,072 became 262,144OutputZ.ai: GLM 5.3z-ai/glm-5.3Supported parameters: include_reasoning, max_tokens, reasoning, reasoning_effort, response_format, temperature, tool_choice, tools, top_k, top_p became frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, parallel_tool_calls, presence_penalty, reasoning, reasoning_effort, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_pAPIZ.ai: GLM Latest~z-ai/glm-latestContext window: 1,048,576 became 1,310,720ContextZ.ai: GLM Latest~z-ai/glm-latestMax output: 131,072 became 943,718OutputZ.ai: GLM Latest~z-ai/glm-latestPrompt price / token: $1.40 / M tokens became $1.15 / M tokensPriceZ.ai: GLM Latest~z-ai/glm-latestCompletion price / token: $4.40 / M tokens became $3.61 / M tokensPriceZ.ai: GLM Latest~z-ai/glm-latestSupported parameters: include_reasoning, max_tokens, reasoning, reasoning_effort, response_format, temperature, tool_choice, tools, top_k, top_p became frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, parallel_tool_calls, presence_penalty, reasoning, reasoning_effort, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_pAPI

21 August 2026

DeepSeek V4 Flash Latest~deepseek/deepseek-v4-flash-latestMax output: 262,144 became 1,048,576OutputDeepSeek V4 Flash Latest~deepseek/deepseek-v4-flash-latestCompletion price / token: $0.14 / M tokens became $0.18 / M tokensPriceDeepSeek: DeepSeek V4 Flash 0423deepseek/deepseek-v4-flashPrompt price / token: $0.08 / M tokens became $0.08 / M tokensPriceDeepSeek: DeepSeek V4 Flash 0423deepseek/deepseek-v4-flashCompletion price / token: $0.16 / M tokens became $0.16 / M tokensPriceDeepSeek: DeepSeek V4 Flash 0731deepseek/deepseek-v4-flash-0731Max output: 393,216 became 384,000OutputDeepSeek: DeepSeek V4 Flash 0731deepseek/deepseek-v4-flash-0731Prompt price / token: $0.14 / M tokens became $0.08 / M tokensPriceDeepSeek: DeepSeek V4 Flash 0731deepseek/deepseek-v4-flash-0731Completion price / token: $0.28 / M tokens became $0.18 / M tokensPriceGoogle: Gemma 4 31Bgoogle/gemma-4-31b-itMax output: 16,384 became 262,144OutputGoogle: Gemma 4 31Bgoogle/gemma-4-31b-itPrompt price / token: $0.09 / M tokens became $0.10 / M tokensPriceLiquidAI: LFM2.5-2.6B (free)liquid/lfm-2.5-2.6b:freeContext window: 128,000 became 65,536ContextMeta: Muse Glimmer 30Bmeta/muse-glimmer-30bMax output: none became 131,072OutputMeta: Muse Glimmer 30Bmeta/muse-glimmer-30bPrompt price / token: $0.35 / M tokens became $0.30 / M tokensPriceMeta: Muse Glimmer 30Bmeta/muse-glimmer-30bCompletion price / token: $1.50 / M tokens became $1.10 / M tokensPriceMiniMax: MiniMax M2.5minimax/minimax-m2.5Max output: none became 32,768OutputMiniMax: MiniMax M2.5minimax/minimax-m2.5Prompt price / token: $0.22 / M tokens became $0.27 / M tokensPriceMiniMax: MiniMax M2.5minimax/minimax-m2.5Completion price / token: $0.90 / M tokens became $0.95 / M tokensPriceMoonshotAI: Kimi K2.6moonshotai/kimi-k2.6Prompt price / token: $0.54 / M tokens became $0.58 / M tokensPriceMoonshotAI: Kimi K2.6moonshotai/kimi-k2.6Completion price / token: $2.28 / M tokens became $2.44 / M tokensPriceMoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-codePrompt price / token: $0.71 / M tokens became $0.67 / M tokensPriceMoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-codeCompletion price / token: $3.50 / M tokens became $3.40 / M tokensPriceQwen: Qwen3.8 2.4T A95Bqwen/qwen3.8-2.4t-a95bMax output: 262,144 became 131,072Output

Sourced from the frozen catalogue snapshots of 20 August 2026 and 21 August 2026 and 3 September 2026. As more snapshots freeze, this timeline grows one dated interval at a time.

New in the catalogue

Entries that appeared between the snapshots of 3 September 2026 and 18 September 2026. Catalogue appearance is not a release date: a model can join the catalogue long after it shipped.
Union Alphastealth/union-alphaLikely Llama3 family · moderate confidenceNewDeepSeek: DeepSeek Pro Latest~deepseek/deepseek-pro-latestlabel: RouterNewDeepSeek: DeepSeek Flash Latest~deepseek/deepseek-flash-latestlabel: RouterNewInference.net: Schematron V2 Turboinference-net/schematron-v2-turbono family on the label yetNewInference.net: Schematron V2 Smallinference-net/schematron-v2-smallno family on the label yetNewOpenAI: GPT Astra Latest~openai/gpt-astra-latestlabel: RouterNewOpenAI: GPT Sol Latest~openai/gpt-sol-latestlabel: RouterNewOpenAI: GPT Terra Latest~openai/gpt-terra-latestlabel: RouterNewOpenAI: GPT Luna Latest~openai/gpt-luna-latestlabel: RouterNewSakana: Fugu Ultra v2sakana/fugu-ultra-v2no family on the label yetNewSakana: Fugu Maxsakana/fugu-maxno family on the label yetNewinclusionAI: Ling 3.0 Flash VLinclusionai/ling-3.0-flash-vlno family on the label yetNewinclusionAI: Ling 3.0 Flash VL (free)inclusionai/ling-3.0-flash-vl:freeno family on the label yetNewDeepSeek: DeepSeek V4.1 Flashdeepseek/deepseek-v4.1-flashlabel: DeepSeekNewInception: Mercury 2.5inception/mercury-2.5no family on the label yetNewNex AGI: Nex-N2.5-Mini (free)nex-agi/nex-n2.5-mini:freelabel: Qwen3NewNex AGI: Nex-N2.5-Pro (free)nex-agi/nex-n2.5-pro:freelabel: Qwen3NewOpenAI: GPT-6 Astraopenai/gpt-6-astralabel: GPTNewOpenAI: GPT-6 Astra (batch)openai/gpt-6-astra:batchlabel: GPTNewOpenAI: GPT-6 Astra Proopenai/gpt-6-astra-prolabel: GPTNewOpenAI: GPT-6 Astra Pro (batch)openai/gpt-6-astra-pro:batchlabel: GPTNewinclusionAI: Ling 3.0 Flash Sante (free)inclusionai/ling-3.0-flash-sante:freeno family on the label yetNewQwen: Qwen3.8 Max (0902)qwen/qwen3.8-max-0902label: QwenNewDeepSeek: DeepSeek V4 Flash Vision Exp (batch)deepseek/deepseek-v4-flash-vision-exp:batchIts own DeepSeek tokenizer group · high confidenceNewZ.ai: GLM 5.3 (batch)z-ai/glm-5.3:batchno family on the label yetNewZ.ai: GLM 5.2 (batch)z-ai/glm-5.2:batchno family on the label yetNewNVIDIA: Nemotron 3.5 Content Safetynvidia/nemotron-3.5-content-safetyno family on the label yetNewSpaceXAI: Grok 4.3 (batch)x-ai/grok-4.3:batchBuilt on the Grok tokenizer · high confidenceNewMistral: Mistral Small 4 (batch)mistralai/mistral-small-2603:batchBuilt on the Mistral tokenizer · high confidenceNewMistral: Ministral 3 8B 2512 (batch)mistralai/ministral-8b-2512:batchBuilt on the Mistral tokenizer · very high confidenceNewMistral: Mistral Large 3 2512 (batch)mistralai/mistral-large-2512:batchBuilt on the Mistral tokenizer · high confidenceNewMistral: Mistral Medium 3.1 (batch)mistralai/mistral-medium-3.1:batchBuilt on the Mistral tokenizer · very high confidenceNewMistral: Codestral 2508 (batch)mistralai/codestral-2508:batchBuilt on the Mistral tokenizer · high confidenceNewTypeSafe: Jev 1.13typesafe/jev-1.13no family on the label yetNew

Gone from the catalogue

41 entries the catalogue no longer lists. Their records and measurements stay published: a model disappearing is a change worth keeping.
Inception: Mercury 2.5 Previewinception/mercury-2.5-previewnever measured before it went · last listed 16 September 2026WithdrawnQwen: Qwen3.8 Maxqwen/qwen3.8-maxmeasured as built on the qwen tokenizer · last listed 16 September 2026WithdrawnThinking Machines: Inkling Small (batch)thinkingmachines/inkling-small:batchmeasured as built on the gpt tokenizer · last listed 16 September 2026WithdrawnNex AGI: Nex-N2-Mininex-agi/nex-n2-minimeasured as built on the qwen tokenizer · last listed 16 September 2026WithdrawnNex AGI: Nex-N2-Pronex-agi/nex-n2-promeasured as built on the qwen tokenizer · last listed 16 September 2026WithdrawnMiniMax: MiniMax M3 (free)minimax/minimax-m3:freemeasured as its own minimax tokenizer group · last listed 16 September 2026WithdrawnIBM: Granite 4.1 8Bibm-granite/granite-4.1-8bmeasured as built on the openai cl100k tokenizer · last listed 16 September 2026WithdrawnOpenAI GPT Latest~openai/gpt-latestnever measured before it went · last listed 16 September 2026WithdrawnGoogle: Gemma 4 31B (batch)google/gemma-4-31b-it:batchmeasured as built on the gemma tokenizer · last listed 16 September 2026WithdrawnMiniMax: MiniMax M2.7 (free)minimax/minimax-m2.7:freemeasured as its own minimax tokenizer group · last listed 16 September 2026WithdrawnNous: Hermes 4 70Bnousresearch/hermes-4-70bmeasured as built on the llama3 tokenizer · last listed 16 September 2026WithdrawnOpenAI: gpt-oss-20b (batch)openai/gpt-oss-20b:batchmeasured as built on the gpt tokenizer · last listed 16 September 2026WithdrawnGoogle: Gemini 2.5 Pro Preview 05-06google/gemini-2.5-pro-preview-05-06measured as built on the gemini tokenizer · last listed 16 September 2026WithdrawnOpenAI: GPT-4 Turbo Previewopenai/gpt-4-turbo-previewnever measured before it went · last listed 16 September 2026WithdrawnOx Alphastealth/ox-alphanever measured before it went · last listed 3 September 2026WithdrawnClaude Opus 5 (Fast)anthropic/claude-opus-5-fastnever measured before it went · last listed 3 September 2026WithdrawnKwaipilot: KAT-Coder-Air V2.5kwaipilot/kat-coder-air-v2.5measured as built on the qwen tokenizer · last listed 3 September 2026WithdrawnMoonshotAI: Kimi K2.7 Code (batch)moonshotai/kimi-k2.7-code:batchmeasured as its own moonshot ai tokenizer group · last listed 3 September 2026WithdrawnNVIDIA: Nemotron 3 Ultra (batch)nvidia/nemotron-3-ultra-550b-a55b:batchmeasured as built on the mistral tokenizer · last listed 3 September 2026WithdrawnAnthropic: Claude Opus 4.8 (Fast)anthropic/claude-opus-4.8-fastnever measured before it went · last listed 3 September 2026WithdrawnAnthropic: Claude Opus 4.7 (Fast)anthropic/claude-opus-4.7-fastnever measured before it went · last listed 3 September 2026WithdrawninclusionAI: Ring-2.6-1Tinclusionai/ring-2.6-1tmeasured as its own inclusionai tokenizer group · last listed 3 September 2026WithdrawninclusionAI: Ling-2.6-1Tinclusionai/ling-2.6-1tmeasured as its own inclusionai tokenizer group · last listed 3 September 2026WithdrawninclusionAI: Ling-2.6-flashinclusionai/ling-2.6-flashmeasured as its own inclusionai tokenizer group · last listed 3 September 2026WithdrawnNVIDIA: Nemotron 3 Nano 30B A3B (free)nvidia/nemotron-3-nano-30b-a3b:freemeasured as built on the mistral tokenizer · last listed 3 September 2026WithdrawnAllenAI: Olmo 3 32B Thinkallenai/olmo-3-32b-thinknever measured before it went · last listed 3 September 2026WithdrawnDeep Cogito: Cogito v2.1 671Bdeepcogito/cogito-v2.1-671bnever measured before it went · last listed 3 September 2026WithdrawnNVIDIA: Nemotron Nano 12B 2 VL (free)nvidia/nemotron-nano-12b-v2-vl:freemeasured as built on the mistral tokenizer · last listed 3 September 2026WithdrawnOpenAI: GPT-5 Codex (batch)openai/gpt-5-codex:batchnever measured before it went · last listed 3 September 2026WithdrawnQwen: Qwen Plus 0728 (thinking)qwen/qwen-plus-2025-07-28:thinkingmeasured as built on the qwen tokenizer · last listed 3 September 2026WithdrawnNVIDIA: Nemotron Nano 9B V2 (free)nvidia/nemotron-nano-9b-v2:freemeasured as built on the mistral tokenizer · last listed 3 September 2026WithdrawnOpenAI: gpt-oss-20b (free)openai/gpt-oss-20b:freemeasured as built on the gpt tokenizer · last listed 3 September 2026WithdrawnOpenAI: o3 Pro (batch)openai/o3-pro:batchnever measured before it went · last listed 3 September 2026WithdrawnGoogle: Gemma 3n 4Bgoogle/gemma-3n-e4b-itmeasured as built on the gemini tokenizer · last listed 3 September 2026WithdrawnArcee AI: Virtuoso Largearcee-ai/virtuoso-largenever measured before it went · last listed 3 September 2026WithdrawnOpenAI: o4 Mini High (batch)openai/o4-mini-high:batchmeasured as built on the gpt tokenizer · last listed 3 September 2026WithdrawnOpenAI: o1-pro (batch)openai/o1-pro:batchnever measured before it went · last listed 3 September 2026WithdrawnOpenAI: o3 Mini High (batch)openai/o3-mini-high:batchmeasured as built on the gpt tokenizer · last listed 3 September 2026WithdrawnOpenAI: o1 (batch)openai/o1:batchnever measured before it went · last listed 3 September 2026WithdrawnMistral: Ministral 8Bmistralai/ministral-8bnever measured before it went · last listed 3 September 2026WithdrawnTheDrummer: Rocinante 12Bthedrummer/rocinante-12bnever measured before it went · last listed 3 September 2026Withdrawn

Stealth watch

Undeclared models, measured while their identity stays hidden
Union Alphastealth/union-alphaLlama3 family candidateStealth

New lineage placements

Models whose catalogue entry names no family, placed for the first time by measurement
inclusionAI: Ling 3.0 Flash Fininclusionai/ling-3.0-flash-finIts own InclusionAI tokenizer groupPlacedinclusionAI: Ling 3.0 Flash Fin (free)inclusionai/ling-3.0-flash-fin:freeIts own InclusionAI tokenizer groupPlacedTencent: Hy-MT2-1.8Btencent/hy-mt2-1.8bIts own Tencent tokenizer groupPlacedTencent: Hy-MT2-30B-A3Btencent/hy-mt2-30b-a3bIts own Tencent tokenizer groupPlacedDots Studio: Dots3-Note Preview (free)dots-studio/dots-3-note-preview:freeBuilt on the Qwen tokenizerPlacedByteDance Seed: Seed 2.1 Turbobytedance-seed/seed-2-1-turboIts own ByteDance Seed tokenizer groupPlacedByteDance Seed: Seed-2.0-Codebytedance-seed/seed-2.0-codeIts own ByteDance Seed tokenizer groupPlacedNVIDIA: Nemotron 3.5 Lightningnvidia/nemotron-3.5-lightningBuilt on the Mistral tokenizerPlacedNVIDIA: Nemotron 3.5 Lightning (free)nvidia/nemotron-3.5-lightning:freeBuilt on the Mistral tokenizerPlacedUpstage: Solar Pro 4upstage/solar-pro4Its own Upstage tokenizer groupPlacedMeta: Muse Glimmer 30Bmeta/muse-glimmer-30bBuilt on the Llama4 tokenizerPlacedMeta: Muse Glimmer 30B (batch)meta/muse-glimmer-30b:batchBuilt on the Llama4 tokenizerPlaced

How the watch works

Each sweep freezes a dated catalogue snapshot and appends new measurements; nothing is overwritten. A new entry is fingerprinted and compared against the whole corpus, a fingerprint change is re-tested before it is published, and a model that leaves the catalogue keeps its record. When identity alerts open, a watched model will be able to notify you; until then, this board and each model's history carry the changes.