AI 模型变更日志 · 2026-06-04

2026-06-04 当日记录 168 个事件。

新增模型 (151)

Alibaba (Qwen)New model: Qwen3.7 Plus
Upstage New model: Solar Pro 3
Anthropic New model: Claude Opus Latest
Anthropic New model: Anthropic Claude Haiku Latest
Anthropic New model: Anthropic Claude Sonnet Latest
Moonshot AI New model: MoonshotAI Kimi Latest
OpenAI New model: GPT Chat Latest
nano-gpt New model: DeepSeek V4 Pro (Thinking)
nano-gpt New model: DeepSeek Latest
nano-gpt New model: Mistral Code Latest
nano-gpt New model: Qwen3.5 27B BlueStar v3 Derestricted Lite
nano-gpt New model: Qwen3.5 27B Marvin DPO V2 Derestricted
nano-gpt New model: Qwen3.5 27B Derestricted
nano-gpt New model: Claude Haiku 4.5 Thinking
nano-gpt New model: Gemma 4 31B K1 v5
nano-gpt New model: Hermes High
nano-gpt New model: Qwen3.5 27B Omega Evolution v2.0 Derestricted Lite
nano-gpt New model: Claw High
nano-gpt New model: Gemma 4 31B Claude 4.6 Opus Reasoning Distilled
nano-gpt New model: Qwen3.5 27B Omega Evolution v2.2 Derestricted Lite
nano-gpt New model: Qwen3.5 27B Vivid Durian
nano-gpt New model: Cohere Command A+ (05/2026)
nano-gpt New model: Qwen3.7 Plus Thinking
nano-gpt New model: Qwen3.5 27B Infracelestial
nano-gpt New model: Qwen3.5 27B NaNovel Derestricted Lite
nano-gpt New model: Qwen3.5 27B Writer V2 Derestricted Lite
nano-gpt New model: Gemma 4 31B Garnet V2
nano-gpt New model: Holo3-35B-A3B Thinking
nano-gpt New model: Qwen3.5 27B Musica v1
nano-gpt New model: Qwen3.5 27B Writer V2 Derestricted
nano-gpt New model: Claw Low
nano-gpt New model: Qwen3.5 27B NaNovel Derestricted
nano-gpt New model: Qwen3.5 27B Queen Derestricted Lite
nano-gpt New model: Gemma 4 31B Larkspur v0.5
nano-gpt New model: Qwen3.5 Flash Thinking
nano-gpt New model: MiroThinker 1.7 Deep Research Mini
nano-gpt New model: Mistral Code Agent Latest
nano-gpt New model: Gemma 4 31B DarkIdol
nano-gpt New model: Hermes Low
nano-gpt New model: Qwen3.5 27B BlueStar Derestricted
nano-gpt New model: Qwen3.5 27B Marvin V2 Derestricted Lite
nano-gpt New model: Hermes Medium
nano-gpt New model: Qwen3.5 27B Marvin DPO V2 Derestricted Lite
nano-gpt New model: Gemma 4 31B MeroMero
nano-gpt New model: Gemma 4 31B Queen
nano-gpt New model: Qwen3.5 27B Queen Derestricted
nano-gpt New model: Mercury Coder Small
nano-gpt New model: Qwen3.5 Omni Plus
nano-gpt New model: MiroThinker 1.7 Deep Research
nano-gpt New model: Claw Medium
nano-gpt New model: Qwen3.5 27B BlueStar v2 Derestricted
nano-gpt New model: Qwen3.5 Omni Flash
nano-gpt New model: Sarvam 105B
nano-gpt New model: Sarvam 30B
nano-gpt New model: ERNIE 5.1 Thinking
nano-gpt New model: Gemma 4 31B Gemopus
nano-gpt New model: Gemma 4 31B Musica v1
nano-gpt New model: Qwen3.5 27B earica Derestricted Lite
nano-gpt New model: Qwen3.5 27B Anko
nano-gpt New model: ERNIE 5.1
nano-gpt New model: Qwen3.5 27B BlueStar v3 Derestricted
nano-gpt New model: Qwen3.7 Max Thinking
nano-gpt New model: Gemma 4 31B Garnet
nano-gpt New model: OWL
nano-gpt New model: Qwen3.5 27B Writer Derestricted
nano-gpt New model: Qwen3.5 122B A10B Thinking
nano-gpt New model: Gemma 4 31B Cognitive Unshackled
nano-gpt New model: Qwen3.5 27B BlueStar v2 Derestricted Lite
nano-gpt New model: Qwen3.5 27B earica Derestricted
nano-gpt New model: Qwen3.5 27B RpRMax v1
nano-gpt New model: Holo3-35B-A3B
nano-gpt New model: Qwen3.5 27B Omega Evolution v2.0 Derestricted
nano-gpt New model: Qwen3.5 27B Writer Derestricted Lite
nano-gpt New model: Gemma 4 31B Fabled
nano-gpt New model: Qwen3.5 35B A3B Thinking
nano-gpt New model: Qwen3.5 27B Thinking
nano-gpt New model: Mercury 2
nano-gpt New model: Qwen3.5 27B BlueStar Derestricted Lite
nano-gpt New model: Qwen3.5 27B Omega Evolution v2.2 Derestricted
nano-gpt New model: Qwen3.5 27B Marvin V2 Derestricted
nano-gpt New model: GLM 5V Turbo Thinking
nano-gpt New model: TheDrummer Skyfall 31B v4.2
nano-gpt New model: Mistral Small 4 119B Thinking
nano-gpt New model: Mistral Small 4 119B
nano-gpt New model: DeepSeek V4 Pro Cheaper
nano-gpt New model: DeepSeek V4 Flash (Thinking)
nano-gpt New model: DeepSeek V4 Pro Cheaper (Thinking)
nano-gpt New model: GLM 4.6 Turbo (Thinking)
nano-gpt New model: GLM 4.7 Original
nano-gpt New model: GLM 4.6V Flash
nano-gpt New model: GLM Latest
nano-gpt New model: GLM 4.7 Flash Original Thinking
nano-gpt New model: GLM 5 Original Thinking
nano-gpt New model: GLM 4.6 Original
nano-gpt New model: GLM 5 Original
nano-gpt New model: GLM 4.6 Turbo
nano-gpt New model: GLM 4.5 (Thinking)
nano-gpt New model: GLM 4.7 Flash Original
nano-gpt New model: GLM 4.7 Original Thinking
nano-gpt New model: GLM 4.7 Thinking

显示 100 / 共 151。

价格变化 (3)

Alibaba (Qwen)Price update: Qwen3.5 Flash now $0.09/$0.36 (was $0.1/$0.4)
Alibaba (Qwen)Price update: Qwen3 Coder 480B A35B now $0.13/$0.5 (was $0.22/$0.95)
nano-gpt Price update: Qwen3.6 35B A3B Thinking now $0.112/$0.8 (was $0.29/$1.74)

上下文变化 (1)

baidu Context window: baidu/ERNIE-4.5-300B-A47B now 123,000 (was 131,072)

输出上限变化 (4)

baidu Max output: baidu/ERNIE-4.5-300B-A47B updated
Mistral Max output: Mistral: Mistral 7B Instruct v0.1 updated
xAI Max output: Grok 4 Fast updated
xAI Max output: x-AI/Grok-Code-Fast 1 updated

能力字段更新 (9)

baidu Capabilities updated: baidu/ERNIE-4.5-300B-A47B
Mistral Capabilities updated: Mistral: Mistral 7B Instruct v0.1
nano-gpt Capabilities updated: Hermes 4 Large
nano-gpt Capabilities updated: Hermes 4 Large (Thinking)
nano-gpt Capabilities updated: Qwen3.6 35B A3B Thinking
xAI Capabilities updated: Grok 3 Beta
xAI Capabilities updated: Grok 3 Mini Beta
xAI Capabilities updated: Grok 4 Fast
xAI Capabilities updated: x-AI/Grok-Code-Fast 1

Pricing and capabilities are refreshed daily and reconciled against each provider's official documentation. Always verify critical production decisions with the provider directly.