Skip to content
AI Atlas
Families

Model families

Families group canonical releases (Llama 3.1, Qwen3, Claude…). Counts, releases and parameter ranges are aggregated from the members' sourced claims; benchmark ranks are the best rank of any member in the benchmark's primary group.

197 families

Reset

2 rows on this page are label-based families (not yet canonical entities)canonical: true = model_family entity; false = legacy attributes.family label awaiting canonicalisation. benchmark_best = best rank of any member in each benchmark's primary group.

Model families
FamilyOrganizationModelsFirst · latest releaseParametersModalitiesLicencesBest benchmark ranks
Nemotron 3NVIDIA74 Dec 2025 · 4 Jun 202631.6BtextOther 1IFBench #4LiveBench Instruction Following #14LiveBench Mathematics #31
Qwen3.8Qwen75 Aug 2026 · 3 Sept 202618.8B – 27.8BimagetextvideoOther 2Apache-2.0 1LiveBench Agentic Coding #5LiveBench Instruction Following #5MMMU-Pro #9
Stable DiffusionStability AI721 Jun 2023 · 29 Oct 202483.7M – 8.15B—Other 4OpenRAIL++-M 2MIT 1no results
GPT 5.4OpenAI65 Mar 2026 · 21 Apr 2026—documentimagetext—Terminal-Bench #6LiveBench Data Analysis #10LiveBench Mathematics #11
gpt-ossOpenAI64 Aug 2025 · 14 May 20261.7B – 116.8BtextApache-2.0 6Aider polyglot #36SWE-bench Verified #36Aider polyglot — well-formed responses #58
GLM4.6Z.ai (Zhipu AI)530 Sept 2025 · 8 Dec 202510.3B – 268.8BimagetextvideoMIT 3SWE-bench Verified #25Terminal-Bench #88τ²-bench #107
GLM4.7Cerebras Systems522 Dec 2025 · 23 Jan 202623B – 358.3BtextMIT 5τ²-bench #3IFBench #73Terminal-Bench #77
GPT 5.1OpenAI513 Nov 2025 · 4 Dec 2025—documentimagetext—SWE-bench Verified #15Terminal-Bench #21MMMU-Pro #39
GPT 5.5OpenAI524 Apr 2026 · 24 Apr 2026—documentimagetext—LiveBench Data Analysis #2Terminal-Bench #4LiveBench Mathematics #6
Gemini 1.5Google5— · ————MMMU-Pro #120GPQA Diamond #284AA Intelligence Index #288
Gemini 3.5Google519 May 2026 · 21 Jul 2026—audiodocumentimagetextvideo—MMMU-Pro #5LiveBench Instruction Following #7LiveBench Language #11
Grok 3SpaceXAI5— · ————Aider polyglot — well-formed responses #1Aider polyglot #26τ²-bench #56
Llama 4Meta AI52 Apr 2025 · 5 Apr 2025108.6BimagetextOther 1Aider polyglot — well-formed responses #17SWE-bench Verified #39Aider polyglot #54
OLMo 3Allen Institute for AI518 Nov 2025 · 19 Nov 20257.3B—Apache-2.0 2IFBench #153Humanity's Last Exam #256Terminal-Bench #256
Phi4Microsoft511 Dec 2024 · 11 Dec 202414.7BtextMIT 1MMMU-Pro #157Terminal-Bench #222GPQA Diamond #295
Qwen3.6Qwen52 Apr 2026 · 27 Apr 202627.8B – 36BimagetextvideoApache-2.0 2τ²-bench #9IFBench #18LiveBench Coding #24
RerankCohere5— · ——text—no results
SonarPerplexity AI527 Jan 2025 · 30 Oct 2025—imagetext—AA Intelligence Index #204Humanity's Last Exam #254GPQA Diamond #266
BERTTogether AI44 Nov 2023 · 13 Nov 202380M—Apache-2.0 3no results
Claude 3Anthropic413 Mar 2024 · 13 Mar 2024—imagetext—SWE-bench (full test split) #1SWE-bench Lite #1MMMU-Pro #150
GLM4.5Z.ai (Zhipu AI)420 Jul 2025 · 20 Oct 202581.9B – 358.3BimagetextMIT 3SWE-bench Verified #27Terminal-Bench #115MMMU-Pro #129
GPTOpenAI4— · ————no results
GPT 3.5OpenAI428 Aug 2023 · 25 Jan 2024—text—SWE-bench (full test split) #6SWE-bench Lite #6AA Intelligence Index #418
GPT 4.1OpenAI414 Apr 2025 · 14 Apr 2025—documentimagetext—Aider polyglot — well-formed responses #1SWE-bench Multimodal #3Aider polyglot #6
GPT 5.2OpenAI410 Dec 2025 · 14 Jan 2026—documentimagetext—LiveBench Coding #4SWE-bench Verified #6SWE-bench Multilingual #9
Grok 4.20xAI431 Mar 2026 · 31 Mar 2026—documentimagetext—IFBench #2τ²-bench #12GPQA Diamond #31
InternLMInternLM (Shanghai AI Laboratory)46 Jul 2023 · 25 Jan 20247B—Apache-2.0 1Other 1no results
Llama 3.3Meta AI426 Nov 2024 · 18 Dec 202470.5BtextLlama-3.3-Community 1IFBench #160τ²-bench #209Terminal-Bench #233
Claude 3.5Anthropic3— · ————Aider polyglot — well-formed responses #1Aider polyglot #15
CodestralMistral AI3— · ————no results
ERNIEBaidu32 Apr 2026 · 18 May 20267.94B – 8.03B—Apache-2.0 3no results
EmbeddingOpenAI316 Jan 2025 · 25 Aug 2026434.1M – 2.72B—CC-BY-NC-4.0 1Other 1no results
GLM5Z.ai (Zhipu AI)311 Feb 2026 · 1 Apr 2026753.9BimagetextvideoMIT 1SWE-bench Multilingual #4τ²-bench #4SWE-bench Verified #6
Granite 4.1IBM36 Apr 2026 · 6 Apr 20263.4B – 28.9B—Apache-2.0 3τ²-bench #160IFBench #171Terminal-Bench #240
Granite 4.2IBM331 Aug 2026 · 31 Aug 2026—text—SciCode #92AA Intelligence Index #160Humanity's Last Exam #179
HunyuanTencent325 Jun 2025 · 30 Jul 20257.5B – 80.4BtextOther 1no results
LFM2Liquid AI312 Aug 2025 · 12 Aug 20251.58B—Other 1IFBench #163Terminal-Bench #279τ²-bench #290
MinistralMistral AI315 Oct 2024 · 15 Oct 20248.02B—Other 1no results
NemotronNVIDIA3— · ————IFBench #7Terminal-Bench #117MMMU-Pro #125
Nemotron 3.5NVIDIA315 May 2026 · 11 Aug 2026638MimagetextOther 1SciCode #108AA Intelligence Index #176GPQA Diamond #192
NorthlabelCohere310 Sept 2026 · 10 Sept 202630B – 218BtextCC-BY-NC-4.0 1no results
Nova 2.0Amazon Web Services3— · ————IFBench #9τ²-bench #45MMMU-Pro #93
OLMo 2.0425Allen Institute for AI317 Apr 2025 · 29 Apr 20251.48B—Apache-2.0 2no results
OLMo 2.1124Allen Institute for AI329 Oct 2024 · 18 Dec 20247.3B—Apache-2.0 2no results
Qwen3.0Qwen327 Apr 2025 · 27 Apr 2025751.6M—Apache-2.0 1τ²-bench #245Terminal-Bench #279Humanity's Last Exam #280
Qwen3.7Qwen321 May 2026 · 27 Jul 2026—imagetextvideo—IFBench #6LiveBench Instruction Following #12Terminal-Bench #12
Seed 2.0ByteDance326 Feb 2026 · 12 Aug 2026—imagetextvideo—no results
Step3.5Cerebras Systems329 Jan 2026 · 25 Feb 2026121B – 149.4BtextOther 2τ²-bench #28Terminal-Bench #74IFBench #83
Yi01-ai323 Dec 2023 · 23 Dec 202334.4B—Apache-2.0 1Aider polyglot — well-formed responses #45Aider polyglot #55
AudiolabelCohere224 Mar 2026 · 18 Jun 20262.07B—Apache-2.0 2no results

Family canonical: true = model_family entity; false = legacy attributes.family label awaiting canonicalisation. benchmark_best = best rank of any member in each benchmark's primary group.