Data health

Which sources were fetched, when, and which model names could not be matched. Unmatched names are listed, never guessed.

Ingestion

Last ingest2026-10-23 07:49Z

Statusup to date

Sources healthy5 / 5

ReleaseLiveBench 2026-10-14 · Terminal-Bench 4.0 · OpenRouter prices · Artificial Analysis

App data built2026-10-23 07:50Z

GitHub Actions re-ingests daily at 02:30 UTC and commits the refreshed stores, which triggers a rebuild of this site.

Per source

OpenRouter catalog (prices)ok

2026-10-23 07:49Z

LiveBench scoresok

2026-10-23 07:49Z

LiveBench costok

2026-10-23 07:49Z

Terminal-Bench leaderboardok

2026-10-23 07:49Z

Artificial Analysis APIok

2026-10-23 07:49Z

Name matching

exact397

unmatched370

manual15

suffix-stripped9

exact = full-name match · manual = reviewed alias table · suffix/date-stripped = matched after removing reasoning-effort tags or run dates · unmatched = never guessed.

Coverage

OpenRouter catalog models499

LiveBench rows matched41 / 49 (8 unmatched)

Terminal-Bench runs matched35 / 35

Artificial Analysis models matched345 / 707 (362 unmatched)

Unmatched rows still appear everywhere in the app, labelled unmatched — they are never coerced onto a similar catalog id, because a wrong match silently corrupts every comparison built on it.

Unmatched LiveBench names (8)

To fix: add these to the alias table in normalization/registry.mjs, then re-run npm run build:data. Each new match raises the coverage numbers above.

Unmatched Terminal-Bench names (0)

Every Terminal-Bench name matched to the catalog.

Unmatched Artificial Analysis names (362)

JT-35B-FlashQwen3.5 Omni PlusNova 2.0 Omni (Medium)Gemma 4 12B (Non-reasoning)Claude 3.5 HaikuGemini 2.0 Flash Thinking Experimental (Jan '25)Qwen2.5 Coder Instruct 32BSonar ReasoningGemma 3 270MGrok 4.20 0309 (Reasoning)Gemini 2.5 Pro Preview (May' 25)o1-previewApodex 1.1Qwen3.5 2B (Non-reasoning)Devstral 2Claude 3.5 Sonnet (Oct '24)Qwen3 Omni 30B A3B InstructGrok Build 0.1 0616Granite 4.2 3BNova 2.0 Pro Preview (Medium)DeepSeek R1 (Jan '25)DeepSeek R1 Distill Llama 8BQwen3 0.6B (Non-reasoning)LFM2 1.2BMagistral Small 1Granite 4.2 30BK2 Horizon 7BLFM 40BClaude 3.7 Sonnet (Reasoning)MiMo-V2-Flash (Feb 2026)Ling-1TGemini 1.5 Pro (May '24)K-EXAONE 2.0 0803Qwen3 4B (Reasoning)Llama 3.3 Nemotron Super 49B v1 (Reasoning)Step 3.5 Flash 2603Mistral Large 2 (Jul '24)DBRX InstructQwen2.5 Instruct 32BNorth Mini CodeMistral MediumNova 2.0 Omni (Non-reasoning)Claude 3 HaikuMistral 7B InstructGrok 4.1 Fast (Reasoning)Hermes 4 - Llama-3.1 70B (Non-reasoning)Grok 3Mistral Small 3.1Llama Nemotron Super 49B v1.5 (Non-reasoning)Grok 2 (Dec '24)MiMo-V2-Omni-0327Qwen3 4B 2507 (Reasoning)Gemini 1.5 Pro (Sep '24)Gemini 1.0 UltraQwen3 1.7B (Non-reasoning)Phi-4 Multimodal InstructGranite 3.3 8B (Non-reasoning)DeepSeek-V2-ChatGrok 4 Fast (Non-reasoning)Ling-2.6-1TNVIDIA Nemotron Nano 9B V2 (Non-reasoning)EXAONE 4.0 32B (Non-reasoning)G9v3-3BQwen3.5 4B (Non-reasoning)K2 Horizon 3.7BOlmo 3 7B ThinkHyperCLOVA X SEED Think (32B)HyperNova 60B 2605 (High, Based on gpt-oss-120b)Nova MicroSarvam M (Reasoning, Based on Mistral Small 3.1)Qwen2.5 TurboEXAONE 4.5 33B (Non-reasoning)Mistral Small (Feb '24)INTELLECT-3 (Based on GLM-4.5-Air)MiniMax M1 40kOlmo 3.1 32B ThinkNex-N2-Pro (Based on Qwen3.5-397B-A17B)LFM2.5-VL-1.6BJT-MINILing-mini-2.0Gemini 2.0 Flash-Lite (Preview)Mixtral 8x7B InstructMolmo2-8BQwen3.5 0.8B (Non-reasoning)Qwen Chat 72BGemini 1.5 Flash (Sep '24)Solar Pro 2 (Preview) (Non-reasoning)MiniCPM-V 4.6 1.3BGPT-4o mini Realtime (Dec '24)Mistral Small 3.2DeepSeek V3 (Dec '24)DeepSeek R1 Distill Qwen 1.5BEXAONE 4.0 32B (Reasoning)Solar Pro 2 (Non-reasoning)Sarvam 105B (High)Llama 3.1 Instruct 405BLlama 3.1 Instruct 70BJT-4.1 Flash 236B A21B (Reasoning)Ring-flash-2.0LFM2.5-1.2B-InstructMistral Large 3Apertus 70B InstructGrok 4.1 Fast (Non-reasoning)Qwen3 0.6B (Reasoning)Grok-1Doubao Seed CodeClaude 4 Opus (Non-reasoning)EXAONE 4.5 33B (Reasoning)Mistral Large 2 (Nov '24)PALM-2Grok 4.20 0309 (Non-reasoning)DeepSeek V3.2 SpecialeGemma 4 E4B (Non-reasoning)K2 Horizon MoVA 36B A4BJamba 1.5 MiniOLMo 2 32BNova ProGranite 4.0 350MGrok 4 Fast (Reasoning)GPT-4o (March 2025, chatgpt-4o-latest)Muse Glimmer (High)Qwen3 1.7B (Reasoning)Kimi Linear 48B A3B InstructLlama 2 Chat 70BSolar Open 100B (Reasoning)LFM2.5-1.2B-ThinkingK2 Think V2Qwen3 Max (Preview)Nova LiteSeed-OSS-36B-InstructGPT-4.5 (Preview)GPT-5 Codex (High)Qwen3.5 2B (Reasoning)MiMo-V2-ProTri-21B-ThinkDeepSeek V4 Pro 0424 (High)Granite 4.0 H SmallQwen3 30B A3B 2507 (Reasoning)Mistral Large (Feb '24)Qwen3.8 MaxGrok 4Llama 3.2 Instruct 11B (Vision)DeepSeek R1 0528 Qwen3 8BLongCat Flash LiteTri-21B-think PreviewNova 2.0 Lite (High)Gemini 2.5 Pro Preview (Mar' 25)MiniCPM5-2BGrok 4.20 0309 v2 (Non-reasoning)Qwen3.5 4B (Reasoning)Magistral Medium 1Motif-2-12.7B-ReasoningMinistral 3 3BLFM2 8B A1BBeam 1.0Llama 2 Chat 13BHermes 4 - Llama-3.1 70B (Reasoning)DeepSeek V4 Pro 0424 (Non-reasoning)NVIDIA Nemotron Nano 12B v2 VL (Reasoning)Qwen3.5 Omni FlashNova 2.0 Lite (Low)Magistral Small 1.2Ministral 3 14BK-EXAONE (Reasoning)Llama 3.2 Instruct 3BGemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning)Llama 3.1 Nemotron Instruct 70BLlama 3.1 Instruct 8BLFM2 2.6BGemma 4 E2B (Non-reasoning)Command-R+ (Apr '24)Sarvam 30B (High)Qwen2 Instruct 72BNova 2.0 Lite (Non-reasoning)DeepSeek V4 Flash Vision (Max)DeepSeek LLM 67B Chat (V1)MiniCPM5-1B (Reasoning)Claude 3 OpusJamba 1.7 LargeGemini 2.0 Flash-Lite (Feb '25)Mistral Small (Sep '24)Nemotron 3.5 Bolt 74B-A9BNanbeige4.1-3BSolar Open2 250BLing 3.0 TinyQwen3 30B A3B 2507 InstructCeleris-1Llama 3.2 Instruct 1BTiny Aya GlobalA.X-K2Exaone 4.0 1.2B (Non-reasoning)Llama 3 Instruct 8BHermes 4 - Llama-3.1 405B (Reasoning)MiMo-V2-Flash (Non-reasoning)Granite 4.1 8BClaude InstantGemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning)Magistral Medium 1.2Gemini 2.5 Flash Preview (Sep '25) (Non-reasoning)QwQ 32B-PreviewLing-flash-2.0Apertus 8B InstructDeepSeek V4 Flash 0420 (High)DeepSeek-V2.5Claude 4 Opus (Reasoning)GPT-4o (Aug '24)Nemotron 3 Nano Omni 30B A3B ReasoningGemini 1.0 ProSolar Pro 2 (Preview) (Reasoning)DeepSeek-Coder-V2Olmo 3 7B InstructJT-4.1 Flash 236B A21B (Non-reasoning)Granite 4.1 3BLing 2.6 FlashJamba 1.6 LargeGPT-6 Sol (Daybreak Blue, max)Gemini 1.5 Flash-8BOlmo 3.1 32B InstructMiMo-V2-OmniMinistral 3 8BLlama 3.3 Instruct 70BNemotron Cascade 2 30B A3BR1 1776Qwen3.8-Flash-NextMiniMax M1 80kLlama 3.2 Instruct 90B (Vision)K-EXAONE (Non-reasoning)Claude 2.1K2 Horizon 0.9BDevstral Small (Jul '25)NVIDIA Nemotron 3 Nano 4BPhi-4 Mini InstructGrok BetaGrok Code Fast 1Nova 2.0 Pro Preview (Non-reasoning)Ring-1TGranite 4.0 1BMotif 3 (Beta)K2-V2 (High)Gemini 3 Deep ThinkERNIE 5.0 Thinking PreviewGranite 4.0 H 350MQwen3.5 0.8B (Reasoning)Apriel-v1.6-15B-ThinkerGemma 3n E4B InstructQwen3 VL 235B A22B (Reasoning)DeepSeek R1 0528 (May '25)Gemma 3n E2B InstructDeepHermes 3 - Mistral 24B Preview (Non-reasoning)Reka Flash (Sep '24)Llama 3 Instruct 70BNVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)Devstral MediumCommand-R (Mar '24)Apriel-v1.5-15B-ThinkerGemini 1.5 Flash (May '24)Claude 3.5 Sonnet (June '24)Step3 VL 10BDeepSeek V4 Flash 0420 (Non-reasoning)ERNIE 4.5 300B A47BMolmo 7B-DMotif 3Qwen3 4B 2507 InstructGemma 4 E2B (Reasoning)Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning)DeepSeek R1 Distill Qwen 14BClaude 3 SonnetGemini 3 Pro Preview (Low)Nova 2.0 Omni (Low)Arctic InstructGPT-4o (ChatGPT)Jamba 1.5 LargeJamba 1.7 MiniQwen2.5 Instruct 72Bo1-miniDeepHermes 3 - Llama-3.1 8B Preview (Non-reasoning)Qwen3 Omni 30B A3B (Reasoning)Qwen2.5 MaxK2 Horizon 375B A23BDeepSeek-V2.5 (Dec '24)Gemini 2.0 Flash (Feb '25)Gemma 3n E4B Instruct Preview (May '25)MiMo-V2-Flash (Reasoning)DeepSeek V4 Pro 0424 (Max)GPT-4o (Nov '24)G9v3-39A5BDiffusionGemma 26B A4BGPT-5.5 Instant (June 2026)Qwen Chat 14BLlama 3.3 Nemotron Super 49B v1 (Non-reasoning)Gemini 2.0 Pro Experimental (Feb '25)Grok 4.20 0309 v2 (Reasoning)DeepSeek Coder V2 Lite InstructGemini 2.5 Flash Preview (Sep '25) (Reasoning)Falcon-H1R-7BOlmo 3 32B ThinkQwen3 Next 80B A3B (Reasoning)Devstral Small (May '25)Qwen3 VL 30B A3B (Reasoning)Devstral Small 2QwQ 32BLlama 3.1 Nemotron Ultra 253B v1 (Reasoning)Gemini 3 Pro Preview (High)Gemma 3 1B InstructQwen1.5 Chat 110BMiniCPM5-1B (Non-reasoning)Qwen3 VL 32B (Reasoning)Hermes 4 - Llama-3.1 405B (Non-reasoning)Granite 4.1 30BExaone 4.0 1.2B (Reasoning)Gemma 4 E4B (Reasoning)DeepSeek V4 Flash 0420 (Max)Claude 3.7 Sonnet (Non-reasoning)Solar MiniK2-V2 (Low)Granite 4.0 H 1BK2-V2 (Medium)Claude 2.0Pixtral LargeQwen3 VL 8B (Reasoning)Phi-3 Mini Instruct 3.8BMi:dm K 2.5 ProDeepSeek R1 Distill Qwen 32BGemini 2.0 Flash Thinking Experimental (Dec '24)Mi:dm K 2.5 Pro PreviewLFM2.5-2.6BLlama Nemotron Super 49B v1.5 (Reasoning)LFM2.5-8B-A1BGPT-4o Realtime (Dec '24)GPT-5.5 Instant (May 2026)Qwen2.5 Coder Instruct 7BQwen3 4B (Non-reasoning)Nova PremierJamba 1.6 MiniGPT-4o (May '24)Muse Wonder Full (Max)Cogito v2.1 (Reasoning)GPT-5 (ChatGPT)Ring-2.6-1TSolar Pro 2 (Reasoning)Muse SparkGrok 3 Reasoning BetaOpenChat 3.5 (1210)NVIDIA Nemotron Nano 9B V2 (Reasoning)Agnes 2.5 Pro BetaNova 2.0 Pro Preview (Low)Qwen3 VL 4B InstructLlama 3.1 Tulu3 405BJamba Reasoning 3BOLMo 2 7BQwen3 VL 4B (Reasoning)Quasar 438B (Max, Based on GLM-5.2)KAT-Coder-Pro V1LFM2 24B A2BNova 2.0 Lite (Medium)KAT Coder Pro V2Gemini 2.0 Flash (experimental)Llama 65BGemma 4 12B (Reasoning)Grok 3 mini Reasoning (High)Llama 2 Chat 7BLing 3.1 Flash 1019

To fix: add these to the manual alias table in normalization/manual-aliases.json, then re-run npm run build:data.