AI model tracker
Every notable AI model release, with its context window, price and official source. Checked daily against the makers' own pages, last updated October 26, 2026.
Latest releases
Oct 24, 2026
Kimi K3.1 ProMoonshot AI (Kimi), China
API · 1M context · $2.00 / $10.00 per 1M
Oct 19, 2026
MiniMax M3.1 FlashMiniMax, China
API · 1M context · $0.15 / $0.60 per 1M
Oct 19, 2026
Muse Glimmer 30BMeta, United States
Open weights
Oct 15, 2026
Nemotron 3.5 Super VLNVIDIA, United States
Open weights · 1M context
| Released | Model | Type | Context | Input / 1M | Output / 1M | Source |
|---|---|---|---|---|---|---|
| Oct 24, 2026 | Kimi K3.1 Pro Moonshot AI (Kimi), China Full weights announced for November 4, 2026. Maximum output tokens not stated by maker. |
API | 1M | $2.00 | $10.00 | kimi.com prices |
| Oct 19, 2026 | MiniMax M3.1 Flash MiniMax, China API model; weights promised later. Maximum output tokens not stated by maker. |
API | 1M | $0.15 | $0.60 | minimax.io |
| Oct 19, 2026 | Muse Glimmer 30B Meta, United States · 29.6B Official Meta research announcement dated Oct 19; model card says August 2026 release. Card states 131,072+ context, without an exact upper limit. |
Open weights Apache 2.0 |
– | – | – | huggingface.co |
| Oct 15, 2026 | Nemotron 3.5 Super VL NVIDIA, United States · 120B (12B active) Open weights; maker states up to 1M context. No published max output or maker token pricing. |
Open weights OpenMDW-1.1 |
1M | – | – | huggingface.co |
| Oct 13, 2026 | Gemini 4 Argon Google, United States Flagship ('Our flagship reasoning model'). Paid tier only (not on free tier). Page announces new prices from Dec 1, 2026: $4.00 in / $0.20 caching / $20.00 out. No >200k tier listed. Context window is a unified input+output limit of 1,048,576; default sync output 65,536, up to 1,048,576 with long... |
API | 1.05M | $2.00 | $10.00 | blog.google prices |
| Oct 12, 2026 | Muse Wonder Meta, United States App model; Meta does not publish token limits or per-token API pricing. |
API | – | – | – | ai.meta.com |
| Oct 12, 2026 | Muse Wonder Full Meta, United States App model; Meta does not publish token limits or per-token API pricing. |
API | – | – | – | ai.meta.com |
| Oct 7, 2026 | Claude Haiku 5.5 Anthropic, United States Small. Prices are for prompts up to 100,000 tokens; prompts over 100,000 tokens: $0.50 in / $0.05 cached / $2.50 out (5m write $0.625, 1h write $1). |
API | 1M | $0.10 | $0.50 | anthropic.com prices |
| Oct 6, 2026 | Mistral Large 4 Mistral, France · 1000B (52B active) Public API preview; weights announced for later release. Maker does not state a max output token limit. |
API | 1M | $1.36 | $4.18 | mistral.ai prices |
| Sep 29, 2026 | GPT-6.1 Sol OpenAI, United States Mid option ('Near-Astra performance for complex work at a lower cost'). Above 272K input: $4.00 / $0.20 / $15.00. Context stated as '1.05M', max output '128K tokens'. |
API | 1.05M | $2.00 | $10.00 | openai.com prices |
| Sep 28, 2026 | Claude Sonnet 5.5 Anthropic, United States Mid ('best combination of speed and intelligence'). Cache hits 0.05x input; 5m write $2.50, 1h write $4. Full 1M context at standard pricing. |
API | 1M | $2.00 | $10.00 | anthropic.com prices |
| Sep 22, 2026 | GPT-6 Sol OpenAI, United States Standard tier. Prompts above 272K input tokens use higher rates for the full request. |
API | 1.05M | $2.00 | $10.00 | openai.com prices |
| Sep 22, 2026 | GPT-6 Luna OpenAI, United States Small ('Our most efficient model for focused, high-volume tasks'). Above 272K input: $0.20 / $0.02 / $0.75. Context stated as '1.05M', max output '128K tokens'. |
API | 1.05M | $0.10 | $0.50 | openai.com prices |
| Sep 22, 2026 | Claude Opus 5.5 Anthropic, United States Anthropic's recommended default ('start with Claude Opus 5.5 for most workloads'). Cache hits 0.05x input; 5m write $5, 1h write $8. Full 1M context at standard pricing. |
API | 1M | $4.00 | $20.00 | anthropic.com prices |
| Sep 21, 2026 | Grok 4.7 xAI, United States No text output limit stated by maker; max_output_tokens remains null. Higher prices apply at 200K+ prompt tokens. |
API | 500K | $2.00 | $6.00 | x.ai prices |
| Sep 15, 2026 | Gemini 3.8 Live Google, United States Live audio/text model; token prices vary by modality, so no single per-token price applies. Caching unsupported. |
API | 131K | – | – | ai.google.dev |
| Sep 10, 2026 | DeepSeek-V4.1-Flash DeepSeek, China Small/fast. PEAK rates; off-peak half: $0.15 / $0.003 / $0.60. Legacy names deepseek-v4-flash and deepseek-v4-flash-vision-exp are routed here. Context stated as '1M', max output 'Up to 384K'. |
API | 1M | $0.30 | $1.20 | api-docs.deepseek.com prices |
| Sep 3, 2026 | GPT-6 Astra OpenAI, United States Flagship ('Our most capable model'). Prices for <=272K input tokens; above 272K: $20.00 in / $2.00 cached / $75.00 out. Context stated as '1.05M', max output as '128K tokens'. Context/max output from developers.openai.com/api/docs/models. |
API | 1.05M | $10.00 | $50.00 | openai.com prices |
| Sep 2, 2026 | Gemini 3.8 Flash Google, United States Mid. Free tier available. Page announces new prices from Jan 1, 2027: $1.50 in / $0.15 caching / $7.50 out. Context caching also has a storage charge. Limits from ai.google.dev/gemini-api/docs/models/gemini-3.8-flash. |
API | 1.05M | $0.75 | $3.75 | ai.google.dev prices |
| Sep 1, 2026 | Claude Fable 5.1 Anthropic, United States Latest Fable model; public Claude API availability. |
API | 1M | $10.00 | $50.00 | platform.claude.com prices |
| Aug 26, 2026 | GLM-5.3-Flash Zhipu (Z.ai), China · 320B (18B active) Open weights and Z.ai API. Maximum output limit and maker token pricing not verified. |
API and open weights MIT |
1M | – | – | z.ai |
| Aug 14, 2026 | GLM-5.3 Zhipu (Z.ai), China Open weights; 1M context and 128K maximum output in Z.ai docs. |
API and open weights GLM-5.3 License |
1M | $1.40 | $4.40 | z.ai prices |
| Aug 13, 2026 | Gemini 3.7 Flash Google, United States Deprecated; API slug routes to Gemini 3.8 Flash. Standard prices through December 31, 2026. |
API | 1.05M | $0.75 | $3.75 | ai.google.dev prices |
| Aug 13, 2026 | DeepSeek-V4-Pro DeepSeek, China Flagship. Prices are PEAK rates (cache-miss input). Off-peak rates are half: $0.66 in / $0.022 cached / $1.98 out. Peak = 01:00-04:00 and 06:00-10:00 UTC Mon-Fri excluding Chinese public holidays. Context stated as '1M', max output 'Up to 384K'. Only two models are listed, so there is no separate... |
API | 1M | $1.32 | $3.96 | api-docs.deepseek.com prices |
| Aug 12, 2026 | Grok 4.6 xAI, United States No text output limit stated by maker; max_output_tokens remains null. Higher prices apply at 200K+ prompt tokens. |
API | 500K | $2.00 | $6.00 | x.ai prices |
| Aug 11, 2026 | Nemotron 3.5 Lightning NVIDIA, United States · 30B (3B active) Full-precision weights; model card says up to 1M context. No maker max output or token pricing stated. |
Open weights OpenMDW-1.1 |
1M | – | – | huggingface.co |
| Aug 8, 2026 | Qwen3.8-2.4T-A95B Alibaba (Qwen), China · 2400B (95B active) Hugging Face repository first published Aug 8; native context 262,144, extensible to 1,010,000. No maker max output or direct token pricing. |
Open weights Qwen3.8-Max License |
262K | – | – | huggingface.co |
| Aug 3, 2026 | Qwen3.8-Max Alibaba (Qwen), China · 2400B (95B active) Priced in CNY; USD fields intentionally null. Open weights are a separate model variant. |
API | 1M | – | – | qwen.ai |
| Jul 31, 2026 | DeepSeek-V4-Flash DeepSeek, China Legacy name routes to V4.1 Flash; shown rates are current peak rates for that route. |
API | 1M | $0.30 | $1.20 | api-docs.deepseek.com prices |
| Jul 24, 2026 | Claude Opus 5 Anthropic, United States Legacy model still available in Claude API. |
API | 1M | $5.00 | $25.00 | anthropic.com prices |
| Jul 21, 2026 | Gemini 3.6 Flash Google, United States Previous-generation Flash model; standard prices through December 31, 2026. |
API | 1.05M | $0.75 | $3.75 | ai.google.dev prices |
| Jul 21, 2026 | Gemini 3.5 Flash-Lite Google, United States Small. Free tier available. Note: older gemini-2.5-flash-lite ($0.10/$0.40) and gemini-3.1-flash-lite ($0.25/$1.50) are cheaper and still listed. Limits from ai.google.dev/gemini-api/docs/models/gemini-3.5-flash-lite. |
API | 1.05M | $0.30 | $2.50 | ai.google.dev prices |
| Jul 16, 2026 | Mistral OCR 4.1 Mistral, France OCR API billed per 1,000 pages, not per million tokens; token context/output limits not stated. |
API | – | – | – | docs.mistral.ai |
| Jul 16, 2026 | Kimi K3 Moonshot AI (Kimi), China · 2800B (104B active) Open weights and API; maximum output tokens not stated by maker. |
API and open weights Kimi K3 License |
1M | $3.00 | $15.00 | kimi.com prices |
| Jun 17, 2026 | Grok 4.3 xAI, United States Release date is the maker's dated Amazon Bedrock GA announcement; earlier xAI API launch date not stated. Max output not stated. |
API | 1M | $1.25 | $2.50 | x.ai prices |
| Jun 16, 2026 | GLM-5.2 Zhipu (Z.ai), China Open weights under MIT; maker states 1M context and 128K maximum output. |
API and open weights MIT |
1M | $1.40 | $4.40 | z.ai prices |
| Jun 4, 2026 | Nemotron 3 Ultra NVIDIA, United States · 550B (55B active) Largest Nemotron reasoning model; open weights. No maker max output or token pricing stated. |
Open weights OpenMDW-1.1 |
1M | – | – | huggingface.co |
| Jun 1, 2026 | MiniMax M3 MiniMax, China · 428B (23B active) Current standard pricing for inputs up to 512K includes a permanent 50% discount; above 512K: $0.60/$0.12/$2.40. Max output not stated. |
API and open weights MiniMax Community License |
1M | $0.30 | $1.20 | minimax.io prices |
| May 28, 2026 | Claude Opus 4.8 Anthropic, United States Legacy model still available in Claude API. |
API | 1M | $5.00 | $25.00 | anthropic.com prices |
| May 22, 2026 | Mistral Medium 3.5 Mistral, France · 128B Maker does not state max output or cached-input price on model/pricing page. |
API and open weights Modified MIT |
256K | $1.50 | $7.50 | mistral.ai prices |
| May 20, 2026 | Qwen3.7-Max Alibaba (Qwen), China Priced in CNY; USD fields intentionally null. |
API | 1M | – | – | qwen.ai |
| Apr 20, 2026 | Kimi K2.6 Moonshot AI (Kimi), China · 1000B (32B active) Open weights and API; maximum output tokens not stated by maker. |
API and open weights Modified MIT |
256K | $0.95 | $4.00 | kimi.com prices |
| Mar 16, 2026 | Mistral Small 4 Mistral, France · 119B (6.5B active) Maker does not state max output or cached-input price on model/pricing page. |
API and open weights Apache 2.0 |
256K | $0.15 | $0.60 | mistral.ai prices |
| Mar 4, 2026 | Phi-4-reasoning-vision-15B Microsoft, United States · 15B Open weights; model card states 16,384 context. No maker token pricing or max output limit stated. |
Open weights MIT |
16K | – | – | huggingface.co |
| Feb 12, 2026 | GLM-5 Zhipu (Z.ai), China · 744B (40B active) API context 200K per Z.ai docs; open weights under MIT. |
API and open weights MIT |
200K | $1.00 | $3.20 | z.ai prices |
| Jan 25, 2026 | Qwen3-Max-Thinking Alibaba (Qwen), China Thinking mode of qwen3-max-2026-01-23; priced in CNY, so USD fields intentionally null. |
API | 262K | – | – | qwen.ai |
Prices are standard US-dollar rates per million tokens from the maker's own pricing page; "–" means the maker does not publish one (open-weight models are free to download and cost what your hardware or host charges). Want to know what a model costs for your workload? Use the API cost calculator.
Change log
- Oct 26, 2026 Added Muse Glimmer 30B (Meta)
- Oct 26, 2026 Added GLM-5.3-Flash (Zhipu (Z.ai))
- Oct 26, 2026 Updated Qwen3.8-Max (Alibaba (Qwen)): note, api id, context tokens, max output tokens
- Oct 26, 2026 Added Qwen3.8-2.4T-A95B (Alibaba (Qwen))
- Oct 26, 2026 Added Nemotron 3 Ultra (NVIDIA)
- Oct 26, 2026 Added Nemotron 3.5 Lightning (NVIDIA)
- Oct 26, 2026 Added Grok 4.3 (xAI)
- Oct 26, 2026 Updated Qwen3-Max-Thinking (Alibaba (Qwen)): note, api id, context tokens, max output tokens
- Oct 26, 2026 Updated GLM-5 (Zhipu (Z.ai)): license, note, api id, context tokens, max output tokens, input per mtok, output per mtok, cached input per mtok, price source
- Oct 26, 2026 Updated Phi-4-reasoning-vision-15B (Microsoft): license, note, context tokens, source
- Oct 26, 2026 Updated Mistral Small 4 (Mistral): note, active params b
- Oct 26, 2026 Updated Kimi K2.6 (Moonshot AI (Kimi)): license, note, api id, context tokens, params b, active params b, input per mtok, output per mtok, cached input per mtok, price source
- Oct 26, 2026 Updated Qwen3.7-Max (Alibaba (Qwen)): note, api id, context tokens, max output tokens
- Oct 26, 2026 Updated Mistral Medium 3.5 (Mistral): note
- Oct 26, 2026 Updated Claude Opus 4.8 (Anthropic): note, api id, context tokens, max output tokens, input per mtok, output per mtok, cached input per mtok, price source
- Oct 26, 2026 Updated MiniMax M3 (MiniMax): license, note, kind, api id, params b, active params b, input per mtok, output per mtok, cached input per mtok, price source
- Oct 26, 2026 Updated GLM-5.2 (Zhipu (Z.ai)): note, api id, max output tokens, input per mtok, output per mtok, cached input per mtok, price source
- Oct 26, 2026 Updated Kimi K3 (Moonshot AI (Kimi)): license, note, api id, active params b, input per mtok, output per mtok, cached input per mtok, price source
- Oct 26, 2026 Updated Mistral OCR 4.1 (Mistral): note, source
- Oct 26, 2026 Updated Gemini 3.6 Flash (Google): note, context tokens, max output tokens, input per mtok, output per mtok, cached input per mtok, price source
- Oct 26, 2026 Updated Claude Opus 5 (Anthropic): note, api id, context tokens, max output tokens, input per mtok, output per mtok, cached input per mtok, price source
- Oct 26, 2026 Updated DeepSeek-V4-Flash (DeepSeek): note, api id, context tokens, max output tokens, input per mtok, output per mtok, cached input per mtok, price source
- Oct 26, 2026 Updated Grok 4.6 (xAI): note, api id, context tokens, cached input per mtok, price source
- Oct 26, 2026 Updated Gemini 3.7 Flash (Google): note, context tokens, max output tokens, input per mtok, output per mtok, cached input per mtok, price source
- Oct 26, 2026 Updated GLM-5.3 (Zhipu (Z.ai)): license, note, api id, context tokens, max output tokens, input per mtok, output per mtok, cached input per mtok, price source
- Oct 26, 2026 Updated Gemini 3.8 Live (Google): note, context tokens, max output tokens, source
- Oct 26, 2026 Updated Grok 4.7 (xAI): note
- Oct 26, 2026 Updated GPT-6 Sol (OpenAI): note, context tokens, max output tokens, cached input per mtok, price source
- Oct 26, 2026 Updated Mistral Large 4 (Mistral): note
- Oct 26, 2026 Updated Muse Wonder Full (Meta): note
How the tracker works
Each entry comes from the maker’s own announcement, documentation, pricing page or model card, and links to it. Prices are standard rates in US dollars per million tokens. When a maker does not publish a figure, the table shows a dash instead of an estimate. Entries are checked daily, and every addition or change is listed in the change log with its date.
What counts as a release
New models from the major AI labs in the United States, China and Europe that anyone can use, through an API, an app or downloadable weights. Third-party fine-tunes, private betas and rumours are left out. Public previews are included and marked as previews.