ainotis Join
My notis

Models

Every model our stories named. Each number links to its story and source. A dot means not checked yet.

All makers AlibabaAnthropicCanvaCloudflareElevenLabsGoogleH2O.aiLiquid AIMetaMicrosoftMicrosoft AIMiniMaxMistralMoonshot AINvidiaOpenAIReflection AISpaceXSIZ.ai
ModelMakerLatestPer million tokens, in / outContext Maker's own benchmarksMeasured by others
Gemini Google 13 Nov not checked yet not checked yet not checked yet not checked yet
Gemini 4 Hydrogen Google 13 Nov $1.50 / $7.50 1,000,000 tokens Terminal-Bench 4.1, default high thinking level 57% · Artificial Analysis Intelligence Index, xhigh thinking level 54 points · Artificial Analysis Intelligence Index, default high thinking level 51 points not checked yet
LaMDA Google 17 Oct not checked yet not checked yet not checked yet not checked yet
Gemma Google 15 Oct not checked yet not checked yet not checked yet not checked yet
gemini-4-argon Google 15 Oct $4 / $20 not checked yet ERQA (Embodied Reasoning), pure reasoning 87% · Gray Swan Indirect Prompt Injection, attack success across 15 adaptive attempts (lower is better) 0.7% · shortcut or cheat attempts on coding benchmarks (lower is better) 2% · LVBench 91.7% · AutomationBench 51.3% · DeepSWE v1.1 77.9% not checked yet
Nano Banana Google 7 Oct not checked yet not checked yet not checked yet not checked yet
Gemini 2.5 Pro Google not checked yet not checked yet not checked yet SHADE-Arena monitoring AUC, benign vs malign transcripts 0.87% · SHADE-Arena overall sabotage success, Claude 3.6 Sonnet as monitor 15%
Gemini 3.8 Flash Google $1.50 / $7.50 not checked yet Terminal-Bench 4.1, as reported by Google in the Gemini 4 Hydrogen launch 18.2% CheatBench cheating rate with Gemini CLI (lower is better) 75.2%
Gemini 4 Carbon Google $4 / $20 not checked yet Gray Swan Indirect Prompt Injection Benchmark attack success rate 0% · Terminal-Bench 4.0 successful cheating 0% · WebDev Arena 1,768 points · CWE-bench 71% · ExploitGym 54.8% · SEC-bench Pro 88.2% · Artificial Analysis Intelligence Index v5.0 59 points not checked yet
Gemini Decision 1 Google $0.10 / not checked yet not checked yet not checked yet not checked yet
Nano Banana 2 Google $0.50 / $60 not checked yet not checked yet not checked yet
Nano Banana 2.1 Google $1.50 / $30 not checked yet not checked yet not checked yet

Maker's own: what a company measured on its own model, from its model card, system card or announcement. A claim about its own product, and labelled so.

Measured by others: a safety institute, a research group or a public benchmark run by someone other than the maker.

Prices for prompts over 100,000 tokens where a model has two tiers. Benchmarks come from each release's own documents, collected by the registrar.

Models

Every model our stories named. Each number links to its story and source. A dot means not checked yet.

All makers AlibabaAnthropicCanvaCloudflareElevenLabsGoogleH2O.aiLiquid AIMetaMicrosoftMicrosoft AIMiniMaxMistralMoonshot AINvidiaOpenAIReflection AISpaceXSIZ.ai

ModelMakerLatest in our stories Price per million tokens, in / outContext windowIndependent test results
Gemini Google New model 13 Nov · · ·
Gemini 4 Hydrogen Google New model 13 Nov $1.50 / $7.50 1,000,000 tokens ·
LaMDA Google 17 Oct · · ·
Gemma Google 15 Oct · · ·
gemini-4-argon Google 15 Oct $4 / $20 · ·
Nano Banana Google 7 Oct · · ·
Gemini 2.5 Pro Google · · SHADE-Arena monitoring AUC, benign vs malign transcripts 0.87% · SHADE-Arena overall sabotage success, Claude 3.6 Sonnet as monitor 15%
Gemini 3.8 Flash Google $1.50 / $7.50 · CheatBench cheating rate with Gemini CLI (lower is better) 75.2%
Gemini 4 Carbon Google $4 / $20 · ·
Gemini Decision 1 Google $0.10 / · · ·
Nano Banana 2 Google $0.50 / $60 · ·
Nano Banana 2.1 Google $1.50 / $30 · ·