Home › Model obsolescence How fast do LLM models become obsolete? A model's real shelf life is set by its own maker's release cadence. Across 207 dated models from 15 labs, the busiest brands ship a new model every few weeks — and the global release rate keeps climbing. Three views: each lab's timeline, its median time between releases, and the market-wide acceleration. This measures release cadence, not forced retirement — old models often stay served, at falling prices.
Every release, brand by brand One row per lab, one dot per model release. Rows sorted by cadence — the density tells the story. Hover a dot for the model; click for its page.
Jul 2024 2025 Jul 2025 2026 Jul 2026 Alibaba · 50 qwen-plus — 2024-01-25 qwen-2.5-72b — 2024-09-19 qwen-2.5-7b — 2024-10-16 qwen-2.5-coder-32b — 2024-11-11 qwen3-next-80b-a3b — 2024-12-01 qwen2.5-vl-72b — 2025-02-01 qwen3-30b-a3b — 2025-04-28 qwen3-14b — 2025-04-28 qwen3-32b — 2025-04-30 qwen3-8b — 2025-04-30 qwen3-235b-a22b — 2025-07-21 qwen3-coder — 2025-07-22 qwen3-coder-plus — 2025-07-23 qwen3-235b-a22b-thinking — 2025-07-25 qwen3-coder-flash — 2025-07-28 qwen3-30b-a3b-thinking — 2025-07-30 qwen3-coder-30b-a3b — 2025-07-31 qwen3-next-80b-a3b-thinking — 2025-09-11 qwen3-vl-235b-a22b — 2025-09-23 qwen3-vl-235b-a22b-thinking — 2025-09-23 qwen3-max — 2025-09-23 qwen3-omni-30b-a3b — 2025-09-24 qwen3-omni-30b-a3b-thinking — 2025-09-24 qwen3-vl-30b-a3b — 2025-10-04 qwen3-vl-30b-a3b-thinking — 2025-10-06 qwen3-vl-8b — 2025-10-14 qwen3-vl-32b — 2025-10-21 qwen3-max-thinking — 2026-01-23 qwen3-coder-next — 2026-02-03 qwen3.5-397b-a17b — 2026-02-15 qwen3.5-plus — 2026-02-16 qwen3.5-35b-a3b — 2026-02-23 qwen3.5-27b — 2026-02-23 qwen3.5-9b — 2026-02-23 qwen3.5-122b-a10b — 2026-02-23 qwen3.5-flash — 2026-02-23 qwen3.6-plus — 2026-04-02 qwen3.6-35b-a3b — 2026-04-17 qwen3.6-max-preview — 2026-04-20 qwen3.6-27b — 2026-04-22 qwen3.6-flash — 2026-04-27 qwen3.7-max — 2026-05-21 qwen3.7-plus — 2026-06-02 qwen3.7-flash — 2026-07-15 qwen3.8-max — 2026-08-03 qwen3.8-2.4t-a95b — 2026-08-12 qwen3.8-27b — 2026-08-14 qwen3.8-flash — 2026-08-26 qwen3.8-omni-flash — 2026-09-17 qwen3.8-max-prime — 2026-09-23 Other · 30 hermes-3-llama-3.1-405b — 2024-08-16 hermes-3-llama-3.1-70b — 2024-08-18 phi-4 — 2024-12-11 reka-flash-3 — 2025-03-12 ernie-4.5-vl-424b-a47b — 2025-06-30 ernie-4.5-21b-a3b — 2025-06-30 dolphin-mistral-24b-venice-edition — 2025-07-09 baichuan-m2-32b — 2025-08-13 hermes-4-405b — 2025-08-26 devstral — 2025-12-09 reka-edge — 2026-03-20 mimo-v2.5 — 2026-04-22 mimo-v2.5-pro — 2026-04-22 step-3.7-flash — 2026-05-29 hy3 — 2026-07-06 inkling — 2026-07-15 ling-3.0-flash — 2026-07-23 inkling-small — 2026-07-30 muse-glimmer-30b — 2026-08-10 hy4-preview — 2026-08-28 muse-spark-1.3 — 2026-09-02 ling-3.0-flash-fin — 2026-09-03 ling-3.0-flash-vl — 2026-09-04 ling-3.0-flash-sante — 2026-09-04 step-5-preview — 2026-09-16 mimo-v2.6-pro — 2026-09-21 mimo-v2.6-flash — 2026-09-22 ling-3.1-flash — 2026-09-29 muse-wonder — 2026-10-12 muse-spark-open — 2026-10-15 Anthropic · 10 claude-opus-4-1 — 2025-08-05 claude-sonnet-4-5 — 2025-09-29 claude-haiku-4-5 — 2025-10-15 claude-fable-5 — 2026-06-09 claude-sonnet-5 — 2026-06-30 claude-opus-5 — 2026-07-24 claude-fable-5-1 — 2026-09-01 claude-opus-5-5 — 2026-09-22 claude-sonnet-5-5 — 2026-09-28 claude-haiku-5-5 — 2026-10-07 Z.ai · 16 glm-4-32b — 2024-01-01 glm-4.5-air — 2025-07-28 glm-4.5 — 2025-07-28 glm-4.5v — 2025-08-11 glm-4.6 — 2025-09-30 glm-4.6v — 2025-12-08 glm-4.7 — 2025-12-22 glm-4.7-flash — 2026-01-19 glm-5 — 2026-02-12 glm-5-turbo — 2026-03-16 glm-5v-turbo — 2026-04-01 glm-5.1 — 2026-04-07 glm-5.2 — 2026-06-13 glm-5.3 — 2026-08-14 glm-5.3-flash — 2026-08-26 glm-5.3-flashx — 2026-09-18 Meta · 8 llama-3.1-8b — 2024-07-23 llama-3.1-70b — 2024-07-23 llama-3.2-3b — 2024-09-25 llama-3.2-1b — 2024-09-25 llama-3.3-70b — 2024-12-06 llama-4-scout — 2025-04-05 llama-4-maverick — 2025-04-05 llama-guard-4-12b — 2025-04-30 Google · 15 gemma-2-27b — 2024-07-13 gemma-3-27b — 2025-03-12 gemma-3-4b — 2025-03-12 gemma-3-12b — 2025-03-13 gemini-2.5-pro — 2025-06-17 gemini-2.5-flash — 2025-06-17 gemini-3-flash-preview — 2025-12-17 gemini-3.1-pro-preview — 2026-02-19 gemma-4-31b — 2026-04-02 gemma-4-26b-a4b — 2026-04-02 gemini-3.6-flash — 2026-07-21 gemini-3.5-flash-lite — 2026-07-21 gemini-3.7-flash — 2026-08-13 gemini-3.8-flash — 2026-09-02 gemini-4-argon — 2026-09-30 Mistral · 18 mixtral-8x22b — 2024-04-17 mistral-nemo — 2024-07-01 mistral-small-24b — 2025-01-30 mistral-saba — 2025-02-17 mistral-small-3.1-24b — 2025-03-17 mistral-medium — 2025-05-07 mistral-medium-3 — 2025-05-07 mistral-small-3.2-24b — 2025-06-20 voxtral-small-24b — 2025-07-15 codestral — 2025-07-30 mistral-medium-3.1 — 2025-08-13 mistral-large — 2025-12-02 ministral-14b — 2025-12-02 ministral-8b — 2025-12-02 ministral-3b — 2025-12-02 mistral-small-4 — 2026-03-16 mistral-medium-3-5 — 2026-04-30 mistral-large-4 — 2026-10-06 DeepSeek · 15 deepseek-chat — 2024-12-26 deepseek-v3 — 2024-12-26 deepseek-chat-v3 — 2025-03-24 deepseek-r1 — 2025-05-28 deepseek-chat-v3.1 — 2025-08-21 deepseek-v3.1-terminus — 2025-09-22 deepseek-v3.2-exp — 2025-09-29 deepseek-v3.2 — 2025-12-01 deepseek-v4-pro-0423 — 2026-04-24 deepseek-v4-flash-0423 — 2026-04-24 deepseek-v4-flash — 2026-07-31 deepseek-v4 — 2026-08-13 deepseek-v4-flash-vision-exp — 2026-08-21 deepseek-v4.1-flash — 2026-09-10 deepseek-v4.1-flash-1017 — 2026-10-17 OpenAI · 15 gpt-4o — 2024-05-13 gpt-4o-mini — 2024-07-18 gpt-4.1-mini — 2025-04-14 gpt-4.1 — 2025-04-14 gpt-oss-20b — 2025-08-05 gpt-oss-120b — 2025-08-05 gpt-oss-safeguard-20b — 2025-10-29 gpt-5.3-codex — 2026-02-24 gpt-5.6-sol — 2026-07-09 gpt-5.6-terra — 2026-07-09 gpt-5.6-luna — 2026-07-09 gpt-6-astra — 2026-09-03 gpt-6-sol — 2026-09-22 gpt-6-luna — 2026-09-22 gpt-6.1-sol — 2026-09-29 xAI · 3 grok-4.5 — 2026-07-08 grok-4.6 — 2026-08-12 grok-4.7 — 2026-09-21 MiniMax · 9 minimax-01 — 2025-01-15 minimax-m1 — 2025-06-17 minimax-m2 — 2025-10-27 minimax-m2.1 — 2025-12-23 minimax-m2-her — 2026-01-23 minimax-m2.5 — 2026-02-12 minimax-m2.7 — 2026-03-18 minimax-m3 — 2026-06-01 minimax-m3.1-flash — 2026-10-19 Moonshot · 5 kimi-k2-thinking — 2025-11-06 kimi-k2.5 — 2026-01-27 kimi-k2.6 — 2026-04-21 kimi-k2.7-code — 2026-06-12 kimi-k3 — 2026-07-16 NVIDIA · 6 nemotron-3-nano-30b-a3b — 2025-12-15 nemotron-3-super-120b-a12b — 2026-03-11 nemotron-3-ultra-550b-a55b — 2026-06-04 nemotron-3.5-content-safety — 2026-06-04 nemotron-3.5-lightning — 2026-08-11 nemotron-3.5-super-vl — 2026-10-15 Cohere · 5 command-r — 2024-08-30 command-r-plus — 2024-08-30 command-r7b — 2024-12-02 command-a — 2025-03-13 command-a-plus — 2026-09-22 IBM · 2 granite-4.0-h-micro — 2025-10-02 granite-4.2-8b — 2026-08-31 A new model every N days Median time between two releases of the same lab — the shorter the bar, the faster yesterday's model stops being the newest.
Alibaba Alibaba: a new model every 7 days (median over 50 releases) · last gap 6 days every 7 days Other Other: a new model every 9 days (median over 30 releases) · last gap 3 days every 9 days Anthropic Anthropic: a new model every 21 days (median over 10 releases) · last gap 9 days every 21 days Z.ai Z.ai: a new model every 24 days (median over 16 releases) · last gap 23 days every 24 days Meta Meta: a new model every 25 days (median over 8 releases) · last gap 25 days every 25 days Google Google: a new model every 26 days (median over 15 releases) · last gap 28 days every 26 days Mistral Mistral: a new model every 28 days (median over 18 releases) · last gap 159 days every 28 days DeepSeek DeepSeek: a new model every 34 days (median over 15 releases) · last gap 37 days every 34 days OpenAI OpenAI: a new model every 38 days (median over 15 releases) · last gap 7 days every 38 days xAI xAI: a new model every 38 days (median over 3 releases) · last gap 40 days every 38 days MiniMax MiniMax: a new model every 66 days (median over 9 releases) · last gap 140 days every 66 days Moonshot Moonshot: a new model every 67 days (median over 5 releases) · last gap 34 days every 67 days NVIDIA NVIDIA: a new model every 68 days (median over 6 releases) · last gap 65 days every 68 days Cohere Cohere: a new model every 98 days (median over 5 releases) · last gap 558 days every 98 days IBM IBM: a new model every 333 days (median over 2 releases) · last gap 333 days every 333 days Median of the gaps between consecutive dated releases of each brand (15 brands with 2+ dated models).
The release rate is accelerating Model releases per quarter across all brands in the index — the market ships faster every year, so shelf lives shrink for everyone.
26 51 2024-Q1: 2 releases 2 2024 2024-Q2: 2 releases 2 Q2 2024-Q3: 12 releases 12 Q3 2024-Q4: 8 releases 8 Q4 2025-Q1: 11 releases 11 2025 2025-Q2: 18 releases 18 Q2 2025-Q3: 30 releases 30 Q3 2025-Q4: 20 releases 20 Q4 2026-Q1: 21 releases 21 2026 2026-Q2: 25 releases 25 Q2 2026-Q3: 51 releases 51 Q3 2026-Q4: 7 releases 7 Q4 Only models with a full, sourced release date and a live paid price are counted (207 of the index). The current quarter is still filling up. Verified 2026-10-23. See also price by model age, per brand and the falling price of intelligence ·methodology .
FAQ How quickly do LLM API models get replaced? Fast, and accelerating. Across 207 dated models, the fastest-shipping lab (Alibaba) puts out a new model every 7 days on median. Most major labs ship a new model every one to two months — a model is rarely the newest of its brand for more than a quarter.
Is the pace of LLM releases accelerating? Yes. Counting the models in this index by release quarter, 2024-Q1 saw 2 releases while 2026-Q4 saw 7 — the release rate has climbed steadily, so the effective shelf life of any given model keeps shrinking.
Does a new release make the previous model obsolete? Not immediately — older models often stay served and get price cuts, which can make them the better value. This page measures release cadence, not forced retirement: it tells you how long a model stays the newest of its line, not when it disappears.