Benchmark

Terminal-Bench Hard

Agentic coding & terminal use

Categories
coding, tool-use
Openness
unknown
Source
artificial_analysis
Reported scores
432

Reported scores

artificial_analysis

ModelOrganizationReported valueReportedEvidence
Apertus 70B InstructSwiss AI Initiative02025-09-02evidence
Apertus 8B InstructSwiss AI Initiative02025-09-02evidence
Apriel-v1.5-15B-ThinkerServiceNow0.1060606060606062025-09-30evidence
Apriel-v1.6-15B-ThinkerServiceNow0.1439393939393942025-11-25evidence
Claude 3 HaikuAnthropic0.007575757575757582024-03-04evidence
Claude 3.5 HaikuAnthropic0.02272727272727272024-10-22evidence
Claude 3.7 Sonnet (Non-reasoning)Anthropic0.2121212121212122025-02-24evidence
Claude 3.7 Sonnet (Reasoning)Anthropic0.2121212121212122025-02-24evidence
Claude 4 Opus (Reasoning)Anthropic0.3106060606060612025-05-22evidence
Claude 4 Sonnet (Non-reasoning)Anthropic0.2727272727272732025-05-22evidence
Claude 4 Sonnet (Reasoning)Anthropic0.3106060606060612025-05-22evidence
Claude 4.1 Opus (Reasoning)Anthropic0.3433523307159092025-08-05evidence
Claude 4.5 Haiku (Non-reasoning)Anthropic0.2727272727272732025-10-15evidence
Claude 4.5 Haiku (Reasoning)Anthropic0.2727272727272732025-10-15evidence
Claude 4.5 Sonnet (Non-reasoning)Anthropic0.2878787878787882025-09-29evidence
Claude 4.5 Sonnet (Reasoning)Anthropic0.3560606060606062025-09-29evidence
Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)Anthropic0.6287878787878792026-06-09evidence
Claude Opus 4.5 (Non-reasoning)Anthropic0.4090909090909092025-11-24evidence
Claude Opus 4.5 (Reasoning)Anthropic0.469696969696972025-11-24evidence
Claude Opus 4.6 (Adaptive Reasoning, Max Effort)Anthropic0.4621212121212122026-02-05evidence
Claude Opus 4.6 (Non-reasoning, High Effort)Anthropic0.4848484848484852026-02-05evidence
Claude Opus 4.7 (Adaptive Reasoning, Max Effort)Anthropic0.5151515151515152026-04-16evidence
Claude Opus 4.7 (Non-reasoning, High Effort)Anthropic0.5454545454545452026-04-16evidence
Claude Opus 4.8 (Adaptive Reasoning, Max Effort)Anthropic0.5833333333333332026-05-28evidence
Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)Anthropic0.530303030303032026-02-17evidence
Claude Sonnet 4.6 (Non-reasoning, High Effort)Anthropic0.4621212121212122026-02-17evidence
Claude Sonnet 4.6 (Non-reasoning, Low Effort)Anthropic0.4242424242424242026-02-17evidence
Cogito v2.1 (Reasoning)Deep Cogito0.1666666666666672025-11-18evidence
Command ACohere0.007575757575757582025-03-13evidence
Command A+Cohere0.252026-05-20evidence
DeepSeek R1 (Jan '25)DeepSeek0.06060606060606062025-01-20evidence
DeepSeek R1 0528 (May '25)DeepSeek0.1590909090909092025-05-28evidence
DeepSeek R1 0528 Qwen3 8BDeepSeek0.01515151515151522025-05-29evidence
DeepSeek R1 Distill Llama 70BDeepSeek0.01515151515151522025-01-20evidence
DeepSeek V3 (Dec '24)DeepSeek0.06818181818181822024-12-26evidence
DeepSeek V3 0324DeepSeek0.1515151515151522025-03-25evidence
DeepSeek V3.1 (Non-reasoning)DeepSeek0.2424242424242422025-08-21evidence
DeepSeek V3.1 (Reasoning)DeepSeek0.252025-08-21evidence
DeepSeek V3.1 Terminus (Non-reasoning)DeepSeek0.3181818181818182025-09-22evidence
DeepSeek V3.1 Terminus (Reasoning)DeepSeek0.3030303030303032025-09-22evidence
DeepSeek V3.2 (Non-reasoning)DeepSeek0.3257575757575762025-12-01evidence
DeepSeek V3.2 (Reasoning)DeepSeek0.3560606060606062025-12-01evidence
DeepSeek V3.2 Exp (Non-reasoning)DeepSeek0.252025-09-29evidence
DeepSeek V3.2 Exp (Reasoning)DeepSeek0.3106060606060612025-09-29evidence
DeepSeek V3.2 SpecialeDeepSeek0.3484848484848492025-12-01evidence
DeepSeek V4 Flash (Non-reasoning)DeepSeek0.3409090909090912026-04-24evidence
DeepSeek V4 Flash (Reasoning, High Effort)DeepSeek0.3863636363636362026-04-24evidence
DeepSeek V4 Flash (Reasoning, Max Effort)DeepSeek0.3560606060606062026-04-24evidence
DeepSeek V4 Pro (Non-reasoning)DeepSeek0.3636363636363642026-04-24evidence
DeepSeek V4 Pro (Reasoning, High Effort)DeepSeek0.4166666666666672026-04-24evidence
DeepSeek V4 Pro (Reasoning, Max Effort)DeepSeek0.4621212121212122026-04-24evidence
Devstral 2Mistral0.1893939393939392025-12-09evidence
Devstral MediumMistral0.09090909090909092025-07-10evidence
Devstral Small (Jul '25)Mistral0.06060606060606062025-07-10evidence
Devstral Small (May '25)Mistral0.06060606060606062025-05-21evidence
Devstral Small 2Mistral0.1666666666666672025-12-09evidence
Doubao Seed CodeByteDance Seed0.2651515151515152025-11-11evidence
ERNIE 4.5 300B A47BBaidu0.06060606060606062025-06-30evidence
ERNIE 5.0 Thinking PreviewBaidu0.252025-11-13evidence
Exaone 4.0 1.2B (Non-reasoning)LG AI Research02025-07-15evidence
Exaone 4.0 1.2B (Reasoning)LG AI Research02025-07-15evidence
EXAONE 4.0 32B (Non-reasoning)LG AI Research0.01515151515151522025-07-15evidence
EXAONE 4.0 32B (Reasoning)LG AI Research0.03787878787878792025-07-15evidence
EXAONE 4.5 33BLG AI Research0.2045454545454552026-04-09evidence
Falcon-H1R-7BTII UAE0.02272727272727272026-01-04evidence
Gemini 2.0 Flash (Feb '25)Google0.03787878787878792025-02-05evidence
Gemini 2.5 Flash (Non-reasoning)Google0.1212121212121212025-05-20evidence
Gemini 2.5 Flash (Reasoning)Google0.1363636363636362025-05-20evidence
Gemini 2.5 Flash Preview (Sep '25) (Non-reasoning)Google0.1439393939393942025-09-25evidence
Gemini 2.5 Flash Preview (Sep '25) (Reasoning)Google0.1666666666666672025-09-25evidence
Gemini 2.5 Flash-Lite (Non-reasoning)Google0.02272727272727272025-06-17evidence
Gemini 2.5 Flash-Lite (Reasoning)Google0.04545454545454552025-06-17evidence
Gemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning)Google0.07575757575757582025-09-25evidence
Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning)Google0.1287878787878792025-09-08evidence
Gemini 2.5 ProGoogle0.2651515151515152025-06-05evidence
Gemini 3 Flash Preview (Non-reasoning)Google0.3181818181818182025-12-17evidence
Gemini 3 Flash Preview (Reasoning)Google0.3863636363636362025-12-17evidence
Gemini 3 Pro Preview (high)Google0.4166666666666672025-11-18evidence
Gemini 3 Pro Preview (low)Google0.3409090909090912025-11-18evidence
Gemini 3.1 Flash-LiteGoogle0.2424242424242422026-03-03evidence
Gemini 3.1 Pro PreviewGoogle0.5378787878787882026-02-19evidence
Gemini 3.5 Flash (high)Google0.4090909090909092026-05-19evidence
Gemini 3.5 Flash (medium)Google0.3939393939393942026-05-19evidence
Gemini 3.5 Flash (minimal)Google0.4621212121212122026-05-19evidence
Gemma 3 12B InstructGoogle0.007575757575757582025-03-12evidence
Gemma 3 1B InstructGoogle02025-03-13evidence
Gemma 3 270MGoogle02025-08-14evidence
Gemma 3 27B InstructGoogle0.03787878787878792025-03-12evidence
Gemma 3 4B InstructGoogle0.007575757575757582025-03-12evidence
Gemma 3n E2B InstructGoogle0.007575757575757582025-06-26evidence
Gemma 3n E4B InstructGoogle0.02272727272727272025-06-26evidence
Gemma 4 12B (Non-reasoning)Google0.1136363636363642026-06-03evidence
Gemma 4 12B (Reasoning)Google0.1818181818181822026-06-03evidence
Gemma 4 26B A4B (Non-reasoning)Google0.252026-04-02evidence
Gemma 4 26B A4B (Reasoning)Google0.13642026-04-02evidence
Gemma 4 31B (Non-reasoning)Google0.3030303030303032026-04-02evidence
Gemma 4 31B (Reasoning)Google0.3636363636363642026-04-02evidence
Gemma 4 E2B (Non-reasoning)Google0.02272727272727272026-04-02evidence
Gemma 4 E2B (Reasoning)Google0.03030303030303032026-04-02evidence
Gemma 4 E4B (Non-reasoning)Google0.07575757575757582026-04-03evidence
Gemma 4 E4B (Reasoning)Google0.08333333333333332026-04-03evidence
GLM 5V Turbo (Reasoning)Z.ai0.3257575757575762026-04-01evidence
GLM-4.5 (Reasoning)Z.ai0.219696969696972025-07-28evidence
GLM-4.5-AirZ.ai0.2045454545454552025-07-28evidence
GLM-4.5V (Non-reasoning)Z.ai0.06818181818181822025-08-11evidence
GLM-4.5V (Reasoning)Z.ai0.0530303030303032025-08-11evidence
GLM-4.6 (Non-reasoning)Z.ai0.2878787878787882025-09-30evidence
GLM-4.6 (Reasoning)Z.ai0.252025-09-30evidence
GLM-4.6V (Non-reasoning)Z.ai0.03030303030303032025-12-08evidence
GLM-4.6V (Reasoning)Z.ai0.1439393939393942025-12-08evidence
GLM-4.7 (Non-reasoning)Z.ai0.3030303030303032025-12-22evidence
GLM-4.7 (Reasoning)Z.ai0.3181818181818182025-12-22evidence
GLM-4.7-Flash (Non-reasoning)Z.ai0.03787878787878792026-01-19evidence
GLM-4.7-Flash (Reasoning)Z.ai0.219696969696972026-01-19evidence
GLM-5 (Non-reasoning)Z.ai0.3939393939393942026-02-11evidence
GLM-5 (Reasoning)Z.ai0.4318181818181822026-02-11evidence
GLM-5-TurboZ.ai0.3333333333333332026-03-15evidence
GLM-5.1 (Non-reasoning)Z.ai0.3560606060606062026-04-07evidence
GLM-5.1 (Reasoning)Z.ai0.4318181818181822026-04-07evidence
GLM-5.2 (max)Z.ai0.5075757575757582026-06-16evidence
GPT-4.1OpenAI0.1363636363636362025-04-14evidence
GPT-4.1 miniOpenAI0.07575757575757582025-04-14evidence
GPT-4.1 nanoOpenAI0.03787878787878792025-04-14evidence
GPT-4o (Aug '24)OpenAI0.08333333333333332024-08-06evidence
GPT-4o (Nov '24)OpenAI0.08333333333333332024-11-20evidence
GPT-5 (ChatGPT)OpenAI0.1287878787878792025-08-07evidence
GPT-5 (high)OpenAI0.3257575757575762025-08-07evidence
GPT-5 (low)OpenAI0.2651515151515152025-08-07evidence
GPT-5 (medium)OpenAI0.3787878787878792025-08-07evidence
GPT-5 (minimal)OpenAI0.1818181818181822025-08-07evidence
GPT-5 Codex (high)OpenAI0.3787878787878792025-09-23evidence
GPT-5 mini (high)OpenAI0.3333333333333332025-08-07evidence
GPT-5 mini (medium)OpenAI0.2878787878787882025-08-07evidence
GPT-5 mini (minimal)OpenAI0.1439393939393942025-08-07evidence
GPT-5 nano (high)OpenAI0.1212121212121212025-08-07evidence
GPT-5 nano (medium)OpenAI0.1742424242424242025-08-07evidence
GPT-5 nano (minimal)OpenAI0.06818181818181822025-08-07evidence
GPT-5.1 (high)OpenAI0.4545454545454552025-11-13evidence
GPT-5.1 (Non-reasoning)OpenAI0.2272727272727272025-11-13evidence
GPT-5.1 Codex (high)OpenAI0.3484848484848492025-11-13evidence
GPT-5.1 Codex mini (high)OpenAI0.3333333333333332025-11-13evidence
GPT-5.2 (medium)OpenAI0.4318181818181822025-12-11evidence
GPT-5.2 (Non-reasoning)OpenAI0.3181818181818182025-12-11evidence
GPT-5.2 (xhigh)OpenAI0.469696969696972025-12-11evidence
GPT-5.2 Codex (xhigh)OpenAI0.3712121212121212025-12-11evidence
GPT-5.3 Codex (xhigh)OpenAI0.530303030303032026-02-05evidence
GPT-5.4 (low)OpenAI0.4318181818181822026-03-05evidence
GPT-5.4 (Non-reasoning)OpenAI0.3787878787878792026-03-05evidence
GPT-5.4 (xhigh)OpenAI0.5757575757575762026-03-05evidence
GPT-5.4 mini (medium)OpenAI0.3409090909090912026-03-17evidence
GPT-5.4 mini (Non-Reasoning)OpenAI0.1818181818181822026-03-17evidence
GPT-5.4 mini (xhigh)OpenAI0.5227272727272732026-03-17evidence
GPT-5.4 nano (medium)OpenAI0.3333333333333332026-03-17evidence
GPT-5.4 nano (Non-Reasoning)OpenAI0.2424242424242422026-03-17evidence
GPT-5.4 nano (xhigh)OpenAI0.4242424242424242026-03-17evidence
GPT-5.5 (high)OpenAI0.5984848484848492026-04-23evidence
GPT-5.5 (low)OpenAI0.5227272727272732026-04-23evidence
GPT-5.5 (medium)OpenAI0.5757575757575762026-04-23evidence
GPT-5.5 (Non-reasoning)OpenAI0.4924242424242422026-04-23evidence
GPT-5.5 (xhigh)OpenAI0.6060606060606062026-04-23evidence
GPT-5.5 Instant (May 2026)OpenAI0.4242424242424242026-05-05evidence
GPT-5.6 Sol (high)OpenAI0.6212121212121212026-07-09evidence
GPT-5.6 Sol (low)OpenAI0.6060606060606062026-07-09evidence
GPT-5.6 Sol (max)OpenAI0.6590909090909092026-07-09evidence
GPT-5.6 Sol (medium)OpenAI0.6287878787878792026-07-09evidence
GPT-5.6 Sol (xhigh)OpenAI0.6136363636363642026-07-09evidence
GPT-5.6 Terra (high)OpenAI0.5757575757575762026-07-09evidence
GPT-5.6 Terra (low)OpenAI0.4393939393939392026-07-09evidence
GPT-5.6 Terra (max)OpenAI0.5757575757575762026-07-09evidence
GPT-5.6 Terra (xhigh)OpenAI0.6287878787878792026-07-09evidence
gpt-oss-120b (high)OpenAI0.2348484848484852025-08-05evidence
gpt-oss-120b (low)OpenAI0.0530303030303032025-08-05evidence
gpt-oss-20b (high)OpenAI0.1060606060606062025-08-05evidence
gpt-oss-20b (low)OpenAI0.04545454545454552025-08-05evidence
Granite 3.3 8B (Non-reasoning)IBM02025-04-16evidence
Granite 4.0 1BIBM02025-10-28evidence
Granite 4.0 350MIBM02025-10-28evidence
Granite 4.0 H 1BIBM02025-10-28evidence
Granite 4.0 H 350MIBM02025-10-28evidence
Granite 4.0 H SmallIBM0.02272727272727272025-09-22evidence
Granite 4.0 MicroIBM0.01515151515151522025-09-22evidence
Granite 4.1 30BIBM0.02272727272727272026-04-29evidence
Granite 4.1 3BIBM0.02272727272727272026-04-29evidence
Granite 4.1 8BIBM02026-04-29evidence
Grok 3xAI0.1136363636363642025-02-19evidence
Grok 3 mini Reasoning (high)xAI0.1742424242424242025-02-19evidence
Grok 4xAI0.3787878787878792025-07-10evidence
Grok 4 Fast (Non-reasoning)xAI0.1212121212121212025-09-19evidence
Grok 4 Fast (Reasoning)xAI0.1893939393939392025-09-19evidence
Grok 4.1 Fast (Non-reasoning)xAI0.1439393939393942025-11-19evidence
Grok 4.1 Fast (Reasoning)xAI0.2424242424242422025-11-19evidence
Grok 4.20 0309 (Non-reasoning)xAI0.219696969696972026-03-10evidence
Grok 4.20 0309 (Reasoning)xAI0.4090909090909092026-03-10evidence
Grok 4.20 0309 v2 (Non-reasoning)xAI0.1666666666666672026-04-07evidence
Grok 4.20 0309 v2 (Reasoning)xAI0.3787878787878792026-04-07evidence
Grok 4.3 (high)xAI0.3787878787878792026-04-30evidence
Grok 4.3 (low)xAI0.2651515151515152026-04-30evidence
Grok 4.3 (medium)xAI0.3030303030303032026-04-30evidence
Grok 4.3 (Non-reasoning)xAI0.1893939393939392026-04-30evidence
Grok Code Fast 1xAI0.1742424242424242025-08-28evidence
Hermes 4 - Llama-3.1 405B (Non-reasoning)Nous Research0.09848484848484852025-08-27evidence
Hermes 4 - Llama-3.1 405B (Reasoning)Nous Research0.1136363636363642025-08-27evidence
Hermes 4 - Llama-3.1 70B (Non-reasoning)Nous Research02025-08-27evidence
Hermes 4 - Llama-3.1 70B (Reasoning)Nous Research0.04545454545454552025-08-27evidence
Hy3-preview (Non-reasoning)Tencent0.3181818181818182026-04-23evidence
Hy3-preview (Reasoning)Tencent0.3409090909090912026-04-23evidence
HyperCLOVA X SEED Think (32B)Naver0.1212121212121212025-12-26evidence
HyperNova 60B 2605Multiverse Computing0.2348484848484852026-05-26evidence
INTELLECT-3Prime Intellect0.09090909090909092025-11-27evidence
Jamba 1.7 LargeAI21 Labs0.02272727272727272025-07-07evidence
Jamba 1.7 MiniAI21 Labs02025-07-07evidence
Jamba Reasoning 3BAI21 Labs0.007575757575757582025-10-08evidence
JT-35B-FlashChina Mobile0.2878787878787882026-05-14evidence
JT-MINIChina Mobile0.1818181818181822026-04-15evidence
K-EXAONE (Non-reasoning)LG AI Research0.06818181818181822025-12-31evidence
K-EXAONE (Reasoning)LG AI Research0.2272727272727272025-12-31evidence
K2 Think V2MBZUAI Institute of Foundation Models0.06818181818181822025-12-15evidence
K2-V2 (high)MBZUAI Institute of Foundation Models0.09848484848484852025-12-05evidence
K2-V2 (low)MBZUAI Institute of Foundation Models0.04545454545454552025-12-05evidence
K2-V2 (medium)MBZUAI Institute of Foundation Models0.08333333333333332025-12-05evidence
KAT Coder Pro V2KwaiKAT0.4924242424242422026-03-27evidence
KAT-Coder-Pro V1KwaiKAT0.09090909090909092025-11-11evidence
Kimi K2Moonshot AI0.1590909090909092025-07-11evidence
Kimi K2 0905Moonshot AI0.2348484848484852025-09-05evidence
Kimi K2 ThinkingMoonshot AI0.3106060606060612025-11-06evidence
Kimi K2.5 (Non-reasoning)Moonshot AI0.1893939393939392026-01-27evidence
Kimi K2.5 (Reasoning)Moonshot AI0.3484848484848492026-01-27evidence
Kimi K2.6Moonshot AI0.4393939393939392026-04-20evidence
Kimi K2.6 (Non-reasoning)Moonshot AI0.3787878787878792026-04-20evidence
Kimi K2.7 CodeMoonshot AI0.4469696969696972026-06-12evidence
Kimi Linear 48B A3B InstructMoonshot AI0.1136363636363642025-10-30evidence
LFM2 1.2BLiquid AI02025-07-10evidence
LFM2 2.6BLiquid AI0.007575757575757582025-09-23evidence
LFM2 24B A2BLiquid AI02026-02-25evidence
LFM2 8B A1BLiquid AI02025-10-07evidence
LFM2.5-1.2B-InstructLiquid AI02026-01-05evidence
LFM2.5-1.2B-ThinkingLiquid AI02026-01-20evidence
LFM2.5-8B-A1BLiquid AI0.04545454545454552026-05-28evidence
LFM2.5-VL-1.6BLiquid AI02026-01-05evidence
Ling 2.6 FlashInclusionAI0.2121212121212122026-04-21evidence
Ling-1TInclusionAI0.1060606060606062025-10-08evidence
Ling-2.6-1TInclusionAI0.3106060606060612026-04-23evidence
Ling-flash-2.0InclusionAI0.1060606060606062025-09-17evidence
Ling-mini-2.0InclusionAI0.007575757575757582025-09-09evidence
Llama 3 Instruct 70BMeta0.007575757575757582024-04-18evidence
Llama 3 Instruct 8BMeta02024-04-18evidence
Llama 3.1 Instruct 405BMeta0.06818181818181822024-07-23evidence
Llama 3.1 Instruct 70BMeta0.03030303030303032024-07-23evidence
Llama 3.1 Instruct 8BMeta0.007575757575757582024-07-23evidence
Llama 3.1 Nemotron Instruct 70BNVIDIA0.04545454545454552024-10-15evidence
Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)NVIDIA0.02272727272727272025-04-07evidence
Llama 3.2 Instruct 11B (Vision)Meta0.007575757575757582024-09-25evidence
Llama 3.2 Instruct 1BMeta02024-09-25evidence
Llama 3.3 Instruct 70BMeta0.03030303030303032024-12-06evidence
Llama 3.3 Nemotron Super 49B v1 (Non-reasoning)NVIDIA02025-03-18evidence
Llama 3.3 Nemotron Super 49B v1 (Reasoning)NVIDIA02025-03-18evidence
Llama 4 MaverickMeta0.06818181818181822025-04-05evidence
Llama 4 ScoutMeta0.01515151515151522025-04-05evidence
Llama Nemotron Super 49B v1.5 (Non-reasoning)NVIDIA0.03787878787878792025-07-25evidence
Llama Nemotron Super 49B v1.5 (Reasoning)NVIDIA0.0530303030303032025-07-25evidence
LongCat Flash LiteLongCat0.1060606060606062026-01-28evidence
Magistral Medium 1Mistral0.09090909090909092025-06-10evidence
Magistral Medium 1.2Mistral0.1287878787878792025-09-18evidence
Magistral Small 1Mistral0.04545454545454552025-06-10evidence
Magistral Small 1.2Mistral0.04545454545454552025-09-17evidence
Mercury 2Inception0.2651515151515152026-02-20evidence
Mi:dm K 2.5 ProKorea Telecom0.02272727272727272025-12-11evidence
Mi:dm K 2.5 Pro PreviewKorea Telecom0.03030303030303032025-12-11evidence
MiMo-V2-Flash (Feb 2026)Xiaomi0.3106060606060612025-12-16evidence
MiMo-V2-Flash (Non-reasoning)Xiaomi0.2575757575757582025-12-16evidence
MiMo-V2-Flash (Reasoning)Xiaomi0.280303030303032025-12-16evidence
MiMo-V2-OmniXiaomi0.3484848484848492026-03-19evidence
MiMo-V2-Omni-0327Xiaomi0.3560606060606062026-03-27evidence
MiMo-V2-ProXiaomi0.4090909090909092026-03-18evidence
MiMo-V2.5Xiaomi0.4166666666666672026-04-22evidence
MiMo-V2.5-ProXiaomi0.4318181818181822026-04-22evidence
MiMo-V2.5-Pro (Non-reasoning)Xiaomi0.3560606060606062026-04-22evidence
MiniCPM-V 4.6 1.3BOpenBMB02026-05-11evidence
MiniCPM5-1B (Non-reasoning)OpenBMB02026-05-25evidence
MiniCPM5-1B (Reasoning)OpenBMB02026-05-25evidence
MiniMax M1 40kMiniMax0.02272727272727272025-06-17evidence
MiniMax M1 80kMiniMax0.03030303030303032025-06-17evidence
MiniMax-M2MiniMax0.2575757575757582025-10-26evidence
MiniMax-M2.1MiniMax0.2878787878787882025-12-23evidence
MiniMax-M2.5MiniMax0.3484848484848492026-02-12evidence
MiniMax-M2.7MiniMax0.3939393939393942026-03-18evidence
MiniMax-M3MiniMax0.4242424242424242026-06-01evidence
Ministral 3 14BMistral0.04545454545454552025-12-02evidence
Ministral 3 3BMistral02025-12-02evidence
Ministral 3 8BMistral0.04545454545454552025-12-02evidence
Mistral Large 2 (Nov '24)Mistral0.06060606060606062024-11-18evidence
Mistral Large 3Mistral0.1590909090909092025-12-02evidence
Mistral Medium 3Mistral0.03787878787878792025-05-07evidence
Mistral Medium 3.1Mistral0.1060606060606062025-08-12evidence
Mistral Medium 3.5Mistral0.3333333333333332026-04-29evidence
Mistral Small 3.1Mistral0.07575757575757582025-03-17evidence
Mistral Small 3.2Mistral0.06818181818181822025-06-20evidence
Mistral Small 4 (Non-reasoning)Mistral0.1060606060606062026-03-16evidence
Mistral Small 4 (Reasoning)Mistral0.1742424242424242026-03-16evidence
Molmo 7B-DAllen Institute for AI02024-09-25evidence
Molmo2-8BAllen Institute for AI02025-12-11evidence
Motif-2-12.7B-ReasoningMotif Technologies0.03787878787878792025-12-04evidence
Muse SparkMeta0.4545454545454552026-04-08evidence
Nanbeige4.1-3BNanbeige02026-02-11evidence
Nemotron 3 Nano Omni 30B A3B ReasoningNVIDIA0.08333333333333332026-04-29evidence
Nemotron 3 Super 120B A12B (Reasoning)NVIDIA0.2878787878787882026-03-11evidence
Nemotron 3 Ultra 550B A55B (Reasoning)NVIDIA0.3636363636363642026-06-04evidence
Nemotron Cascade 2 30B A3BNVIDIA0.2121212121212122026-03-19evidence
North Mini CodeCohere0.3106060606060612026-06-09evidence
Nova 2.0 Lite (high)Amazon0.1666666666666672025-10-29evidence
Nova 2.0 Lite (low)Amazon0.03787878787878792025-10-29evidence
Nova 2.0 Lite (medium)Amazon0.1742424242424242025-10-29evidence
Nova 2.0 Lite (Non-reasoning)Amazon0.06818181818181822025-10-29evidence
Nova 2.0 Omni (low)Amazon0.03787878787878792025-11-26evidence
Nova 2.0 Omni (medium)Amazon0.04545454545454552025-11-26evidence
Nova 2.0 Omni (Non-reasoning)Amazon0.06818181818181822025-11-26evidence
Nova 2.0 Pro Preview (low)Amazon0.1742424242424242025-11-27evidence
Nova 2.0 Pro Preview (medium)Amazon0.2424242424242422025-11-27evidence
Nova 2.0 Pro Preview (Non-reasoning)Amazon0.1666666666666672025-11-27evidence
Nova LiteAmazon0.007575757575757582024-12-03evidence
Nova MicroAmazon0.01515151515151522024-12-03evidence
Nova PremierAmazon0.06818181818181822025-04-30evidence
Nova ProAmazon0.06060606060606062024-12-03evidence
NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)NVIDIA0.1212121212121212025-12-15evidence
NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)NVIDIA0.1363636363636362025-12-15evidence
NVIDIA Nemotron 3 Nano 4BNVIDIA0.06818181818181822026-03-16evidence
NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)NVIDIA02025-10-28evidence
NVIDIA Nemotron Nano 12B v2 VL (Reasoning)NVIDIA0.04545454545454552025-10-28evidence
NVIDIA Nemotron Nano 9B V2 (Non-reasoning)NVIDIA0.007575757575757582025-08-18evidence
NVIDIA Nemotron Nano 9B V2 (Reasoning)NVIDIA0.01515151515151522025-08-18evidence
o1OpenAI0.1287878787878792024-12-05evidence
o3OpenAI0.3712121212121212025-04-16evidence
o3-miniOpenAI0.06818181818181822025-01-31evidence
o3-mini (high)OpenAI0.06060606060606062025-01-31evidence
o4-mini (high)OpenAI0.1515151515151522025-04-16evidence
OLMo 2 32BAllen Institute for AI02025-03-13evidence
OLMo 2 7BAllen Institute for AI02024-11-26evidence
Olmo 3 32B ThinkAllen Institute for AI0.01515151515151522025-11-20evidence
Olmo 3 7B InstructAllen Institute for AI02025-11-20evidence
Olmo 3 7B ThinkAllen Institute for AI0.007575757575757582025-11-20evidence
Olmo 3.1 32B InstructAllen Institute for AI02026-01-13evidence
Olmo 3.1 32B ThinkAllen Institute for AI02025-12-12evidence
Phi-3 Mini Instruct 3.8BMicrosoft02024-04-23evidence
Phi-4Microsoft0.03787878787878792024-12-12evidence
Phi-4 Mini InstructMicrosoft02024-02-26evidence
Qwen2.5 Instruct 72BQwen0.04545454545454552024-09-19evidence
Qwen3 0.6B (Non-reasoning)Qwen02025-04-28evidence
Qwen3 0.6B (Reasoning)Qwen02025-04-28evidence
Qwen3 1.7B (Non-reasoning)Qwen02025-04-28evidence
Qwen3 1.7B (Reasoning)Qwen02025-04-28evidence
Qwen3 14B (Non-reasoning)Qwen0.0530303030303032025-04-28evidence
Qwen3 14B (Reasoning)Qwen0.03787878787878792025-04-28evidence
Qwen3 235B A22B (Non-reasoning)Qwen0.06060606060606062025-04-28evidence
Qwen3 235B A22B (Reasoning)Qwen0.06060606060606062025-04-28evidence
Qwen3 235B A22B 2507 (Reasoning)Qwen0.1363636363636362025-07-25evidence
Qwen3 235B A22B 2507 InstructQwen0.1515151515151522025-07-21evidence
Qwen3 30B A3B (Non-reasoning)Qwen0.06818181818181822025-04-28evidence
Qwen3 30B A3B (Reasoning)Qwen0.02272727272727272025-04-28evidence
Qwen3 30B A3B 2507 (Reasoning)Qwen0.0530303030303032025-07-30evidence
Qwen3 30B A3B 2507 InstructQwen0.06060606060606062025-07-29evidence
Qwen3 32B (Reasoning)Qwen0.03030303030303032025-04-28evidence
Qwen3 4B 2507 (Reasoning)Qwen0.01515151515151522025-08-06evidence
Qwen3 4B 2507 InstructQwen0.04545454545454552025-08-06evidence
Qwen3 8B (Non-reasoning)Qwen0.02272727272727272025-04-28evidence
Qwen3 8B (Reasoning)Qwen0.02272727272727272025-04-28evidence
Qwen3 Coder 30B A3B InstructQwen0.1515151515151522025-07-31evidence
Qwen3 Coder 480B A35B InstructQwen0.1893939393939392025-07-22evidence
Qwen3 Coder NextQwen0.1818181818181822026-02-03evidence
Qwen3 MaxQwen0.2045454545454552025-09-23evidence
Qwen3 Max (Preview)Qwen0.1969696969696972025-09-05evidence
Qwen3 Max ThinkingQwen0.2424242424242422026-01-26evidence
Qwen3 Max Thinking (Preview)Qwen0.1742424242424242025-11-03evidence
Qwen3 Next 80B A3B (Reasoning)Qwen0.09848484848484852025-09-11evidence
Qwen3 Next 80B A3B InstructQwen0.07575757575757582025-09-11evidence
Qwen3 Omni 30B A3B (Reasoning)Qwen0.03787878787878792025-09-22evidence
Qwen3 Omni 30B A3B InstructQwen0.01515151515151522025-09-22evidence
Qwen3 VL 235B A22B (Reasoning)Qwen0.1136363636363642025-09-23evidence
Qwen3 VL 235B A22B InstructQwen0.06818181818181822025-09-23evidence
Qwen3 VL 30B A3B (Reasoning)Qwen0.0530303030303032025-10-03evidence
Qwen3 VL 30B A3B InstructQwen0.06060606060606062025-10-03evidence
Qwen3 VL 32B (Reasoning)Qwen0.07575757575757582025-10-21evidence
Qwen3 VL 32B InstructQwen0.08333333333333332025-10-21evidence
Qwen3 VL 4B (Reasoning)Qwen0.01515151515151522025-10-14evidence
Qwen3 VL 4B InstructQwen02025-10-14evidence
Qwen3 VL 8B (Reasoning)Qwen0.03787878787878792025-10-14evidence
Qwen3 VL 8B InstructQwen0.02272727272727272025-10-14evidence
Qwen3.5 0.8B (Non-reasoning)Qwen02026-03-02evidence
Qwen3.5 0.8B (Reasoning)Qwen02026-03-02evidence
Qwen3.5 122B A10B (Non-reasoning)Qwen0.2954545454545452026-02-24evidence
Qwen3.5 122B A10B (Reasoning)Qwen0.3106060606060612026-02-24evidence
Qwen3.5 27B (Non-reasoning)Qwen0.3181818181818182026-02-24evidence
Qwen3.5 27B (Reasoning)Qwen0.3257575757575762026-02-24evidence
Qwen3.5 2B (Non-reasoning)Qwen0.03787878787878792026-03-02evidence
Qwen3.5 2B (Reasoning)Qwen0.03787878787878792026-03-02evidence
Qwen3.5 35B A3B (Non-reasoning)Qwen0.1060606060606062026-02-24evidence
Qwen3.5 35B A3B (Reasoning)Qwen0.2651515151515152026-02-24evidence
Qwen3.5 397B A17B (Non-reasoning)Qwen0.3560606060606062026-02-16evidence
Qwen3.5 397B A17B (Reasoning)Qwen0.4090909090909092026-02-16evidence
Qwen3.5 4B (Non-reasoning)Qwen0.1136363636363642026-03-02evidence
Qwen3.5 4B (Reasoning)Qwen0.1818181818181822026-03-02evidence
Qwen3.5 9B (Non-reasoning)Qwen0.1818181818181822026-03-02evidence
Qwen3.5 9B (Reasoning)Qwen0.2424242424242422026-03-02evidence
Qwen3.5 Omni FlashQwen0.08333333333333332026-03-30evidence
Qwen3.5 Omni PlusQwen0.2121212121212122026-03-30evidence
Qwen3.6 27B (Non-reasoning)Qwen0.2121212121212122026-04-22evidence
Qwen3.6 27B (Reasoning)Qwen0.3484848484848492026-04-22evidence
Qwen3.6 35B A3B (Non-reasoning)Qwen0.2575757575757582026-04-16evidence
Qwen3.6 35B A3B (Reasoning)Qwen0.3484848484848492026-04-16evidence
Qwen3.6 Max PreviewQwen0.4393939393939392026-04-20evidence
Qwen3.6 PlusQwen0.4393939393939392026-04-02evidence
Qwen3.7 MaxQwen0.5075757575757582026-05-19evidence
Qwen3.7 PlusQwen0.469696969696972026-06-01evidence
Reka Flash 3Reka AI02025-03-10evidence
Ring-1TInclusionAI0.06818181818181822025-10-13evidence
Ring-2.6-1TInclusionAI0.2878787878787882026-05-08evidence
Ring-flash-2.0InclusionAI0.07575757575757582025-09-19evidence
Sarvam 105B (high)Sarvam0.01515151515151522026-03-06evidence
Sarvam 30B (high)Sarvam0.02272727272727272026-03-06evidence
Sarvam M (Reasoning)Sarvam0.02272727272727272025-05-23evidence
Seed-OSS-36B-InstructByteDance Seed0.06818181818181822025-08-20evidence
Solar Open 100B (Reasoning)Upstage0.02272727272727272025-12-17evidence
Solar Pro 2 (Non-reasoning)Upstage0.04545454545454552025-07-09evidence
Solar Pro 2 (Reasoning)Upstage0.03030303030303032025-07-09evidence
Solar Pro 3Upstage0.07575757575757582026-04-06evidence
Step 3.5 FlashStepFun0.2727272727272732026-02-02evidence
Step 3.5 Flash 2603StepFun0.3257575757575762026-04-02evidence
Step 3.7 FlashStepFun0.3560606060606062026-05-29evidence
Step3 VL 10BStepFun0.0530303030303032026-01-20evidence
Tiny Aya GlobalCohere02026-02-17evidence
Tri-21B-ThinkTrillion Labs0.007575757575757582026-02-10evidence
Tri-21B-think PreviewTrillion Labs0.02272727272727272026-02-10evidence
Trinity Large ThinkingArcee AI0.2272727272727272026-04-01evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.