BlogLLM Benchmark Scores as Published, Side by Side
OSWorld: Results as Published
Every result the catalog holds for OSWorld: 71 models from 12 developers, read from 46 pages. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area Computer UseMaintainer XLANG Lab, University of Hong KongIntroduced 2024Data version 2026103001, as of 2026-10-30
OSWorld tests whether an AI agent can carry out open-ended tasks on a real computer desktop, on Ubuntu, Windows or macOS, by clicking and typing in real applications. OSWorld 2.0 has 108 long workflows across professional and everyday work, each taking a skilled human over an hour, with some changing mid-task and some requiring questions to the user. Success is binary completion within a 500-step limit, with a partial score for checkpoints reached.
OSWorld-Verified
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-5.6 LunaOpenAI, released 2026-07-09 | 72.6% | Google, Gemini 3.6 Flash model cardchecked 2026-10-15 | - |
| GPT-5.6 SolOpenAI, released 2026-07-09 | 83% | Moonshot AI, Kimi K3 model cardchecked 2026-10-15; printed on 1 more page | Effort: max |
| GPT-5.5OpenAI, released 2026-04-23 | 78.7% | OpenAI, Introducing GPT-5.5checked 2026-10-15; printed on 5 more pages | Effort: xhigh |
| GPT-5.4 miniOpenAI, released 2026-03-17 | 72.1% | OpenAI, Introducing GPT-5.4 mini and nanochecked 2026-10-15; printed on 1 more page | Effort: xhigh |
| GPT-5.4 nanoOpenAI, released 2026-03-17 | 39% | OpenAI, Introducing GPT-5.4 mini and nanochecked 2026-10-15 | Effort: xhigh |
| GPT-5.4OpenAI, released 2026-03-05 | 75% | OpenAI, Introducing GPT-5.4checked 2026-10-15; printed on 4 more pages | Effort: xhigh |
| GPT-5.3-CodexOpenAI, released 2026-02-05 | 64.7% | OpenAI, Introducing GPT-5.3-Codexchecked 2026-10-15; printed on 1 more page | Effort: xhigh |
| GPT-5.2-CodexOpenAI, released 2025-12-18 | 38.2% | OpenAI, Introducing GPT-5.3-Codexchecked 2026-10-15 | Effort: xhigh |
| GPT-5.2OpenAI, released 2025-12-11 | 37.9% | OpenAI, Introducing GPT-5.3-Codexchecked 2026-10-15; printed on 3 more pages | Effort: xhigh |
| GPT-5 miniOpenAI, released 2025-08-07 | 42% | OpenAI, Introducing GPT-5.4 mini and nanochecked 2026-10-15 | Effort: high |
| Claude Opus 5Anthropic, released 2026-07-24 | 83.4% | Xiaomi, Xiaomi MiMo-V2.6-Pro-RL model cardchecked 2026-10-19 | - |
| Claude Sonnet 5Anthropic, released 2026-06-30 | 81.2% | Anthropic, Claude Sonnet 5checked 2026-10-15; printed on 1 more page | - |
| Claude Fable 5Anthropic, released 2026-06-09 | 85% | Moonshot AI, Kimi K3 model cardchecked 2026-10-15; printed on 1 more page | Effort: max; Fallback: w/ fallback |
| Claude Mythos 5Anthropic, released 2026-06-09 | 85% | Anthropic, Claude Fable 5 and Claude Mythos 5checked 2026-10-15 | - |
| Claude Opus 4.8Anthropic, released 2026-05-28 | 83.4% | Anthropic, Claude Opus 4.8checked 2026-10-15; printed on 3 more pages | - |
| Claude Opus 4.7Anthropic, released 2026-04-16 | 78% | Anthropic, Claude Opus 4.7checked 2026-10-15; printed on 4 more pages | - |
| Claude Mythos PreviewAnthropic, released 2026-04-07 | 85.4% | Anthropic, Claude Fable 5 and Claude Mythos 5checked 2026-10-15; printed on 1 more page | - |
| Claude Sonnet 4.6Anthropic, released 2026-02-17 | 72.5% | Anthropic, Claude Sonnet 4.6checked 2026-10-15; printed on 3 more pages | - |
| Claude Opus 4.6Anthropic, released 2026-02-05 | 72.7% | Anthropic, Claude Opus 4.6checked 2026-10-15; printed on 4 more pages | - |
| Claude Opus 4.5Anthropic, released 2025-11-24 | 66.3% | Anthropic, Claude Opus 4.5checked 2026-10-15; printed on 3 more pages | Effort: high |
| Claude Haiku 4.5Anthropic, released 2025-10-15 | 50.7% | Anthropic, Claude Haiku 4.5checked 2026-10-15; printed on 1 more page | Effort: thinking budget 128K |
| Claude Sonnet 4.5Anthropic, released 2025-09-29 | 61.4% | Anthropic, Claude Sonnet 4.5checked 2026-10-15; printed on 5 more pages | - |
| Claude Opus 4.1Anthropic, released 2025-08-05 | 44.4% | Anthropic, Claude Opus 4.5checked 2026-10-15; printed on 1 more page | - |
| Claude Sonnet 4Anthropic, released 2025-05-22 | 42.2% | Anthropic, Claude Sonnet 4.5checked 2026-10-15; printed on 2 more pages | - |
| Claude 3.7 SonnetAnthropic, released 2025-02-24 | 28% | Anthropic, Claude Sonnet 4.6checked 2026-10-15 | - |
| Claude 3.5 SonnetAnthropic, released 2024-10-22 | 14.9% | Anthropic, Claude Sonnet 4.6checked 2026-10-15 | - |
| Gemini 3.5 Flash-LiteGoogle, released 2026-07-21 | 74% | Google, Gemini 3.5 Flash-Lite model cardchecked 2026-10-15 | - |
| Gemini 3.6 FlashGoogle, released 2026-07-21 | 83% | Google, Gemini 3.6 Flash model cardchecked 2026-10-15 | - |
| Gemini 3.5 FlashGoogle, released 2026-05-19 | 78.4% | Google, Gemini 3.5 Flash model cardchecked 2026-10-15; printed on 1 more page | - |
| Gemini 3.1 Flash-LiteGoogle, released 2026-03-03 | 54.3% | Google, Gemini 3.5 Flash-Lite model cardchecked 2026-10-15 | - |
| Gemini 3.1 Pro PreviewGoogle, released 2026-02-19 | 76.2% | Google, Gemini 3.5 Flash model cardchecked 2026-10-15; printed on 4 more pages | - |
| Gemini 3 FlashGoogle, released 2025-12-17 | 65.1% | Google, Gemini 3.5 Flash model cardchecked 2026-10-15 | - |
| Muse Spark OpenMeta, released 2026-10-15 | 80.4% | Meta, Meta Muse Spark Open 816B-A42B model cardchecked 2026-10-19 | Effort: xhigh |
| Muse Glimmer-30BMeta, released 2026-08-09 | 65.9% | Alibaba, Qwen3.8-27B model cardchecked 2026-10-15 | - |
| DeepSeek-V4-Pro (Preview)DeepSeek, released 2026-04-24 | 80.6% | MiniMax, MiniMax M3checked 2026-10-15 | - |
| Qwen3.7-PlusAlibaba, released 2026-06-01 | 73.3% | Alibaba, Qwen3.8-27B model cardchecked 2026-10-15 | - |
| Qwen3.6-27BAlibaba, released 2026-04-22 | 63.9% | Alibaba, Qwen3.8-27B model cardchecked 2026-10-15 | - |
| Qwen3.5-122B-A10BAlibaba, released 2026-02-24 | 58% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15; printed on 1 more page | - |
| Qwen3.5-27BAlibaba, released 2026-02-24 | 56.2% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| Qwen3.5-35B-A3BAlibaba, released 2026-02-24 | 54.5% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| Qwen3.5-397B-A17BAlibaba, released 2026-02-16 | 62.2% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| Qwen3-VL-235B-A22BAlibaba, released 2025-09-23 | 38.1% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15; printed on 1 more page | - |
| Qwen3.8-27BAlibaba | 84.3% | Alibaba, Qwen3.8-27B model cardchecked 2026-10-15 | - |
| Qwen3.8-Flash-NextAlibaba | 55.8% | NVIDIA, NVIDIA Nemotron 3.5 Super VL 120B-A12B model cardchecked 2026-10-19 | - |
| Kimi K3Moonshot AI, released 2026-07-16 | 84.8% | Moonshot AI, Kimi K3 model cardchecked 2026-10-15; printed on 1 more page | Effort: max |
| Kimi K2.6Moonshot AI, released 2026-04-20 | 73.1% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 1 more page | - |
| Kimi K2.5Moonshot AI, released 2026-01-27 | 63.3% | Moonshot AI, Kimi K2.6 model cardchecked 2026-10-15; printed on 1 more page | - |
| GLM-5.3-FlashZ.ai, released 2026-08-26 | 14.78% | NVIDIA, NVIDIA Nemotron 3.5 Super VL 120B-A12B model cardchecked 2026-10-19 | - |
| MiniMax M3MiniMax, released 2026-06-01 | 70.06% | MiniMax, MiniMax M3checked 2026-10-15 | - |
| Nemotron 3.5 Super VL 120B-A12BNVIDIA, released 2026-10-15 | 51.14% | NVIDIA, NVIDIA Nemotron 3.5 Super VL 120B-A12B model cardchecked 2026-10-19 | - |
| MiMo-V2.6-FlashXiaomi, released 2026-09-21 | 80.8% | Xiaomi, Xiaomi MiMo-V2.6-Pro-RL model cardchecked 2026-10-19 | - |
| MiMo-V2.6-ProXiaomi, released 2026-09-21 | 82% | Xiaomi, Xiaomi MiMo-V2.6-Pro-RL model cardchecked 2026-10-19 | - |
| InklingThinking Machines, released 2026-07-14 | 40.1% | Meta, Meta Muse Spark Open 816B-A42B model cardchecked 2026-10-19 | Effort: xhigh |
OSWorld 2.0
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-5.6 LunaOpenAI, released 2026-07-09 | 45.6% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| GPT-5.6 SolOpenAI, released 2026-07-09 | 62.6% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 3 more pages | - |
| GPT-5.6 TerraOpenAI, released 2026-07-09 | 50.2% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 1 more page | - |
| GPT-5.5OpenAI, released 2026-04-23 | 47.5% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 1 more page | - |
| Claude Opus 5Anthropic, released 2026-07-24 | 70.6% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 1 more page | - |
| Claude Fable 5Anthropic, released 2026-06-09 | 66.1% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 1 more page | - |
| Claude Opus 4.8Anthropic, released 2026-05-28 | 55.7% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 2 more pages | - |
| Gemini 3.7 FlashGoogle, released 2026-08-13 | 47.9% | Google, Gemini 3.7 Flash model cardchecked 2026-10-15 | - |
| Gemini 3.6 FlashGoogle, released 2026-07-21 | 33.8% | Google, Gemini 3.7 Flash model cardchecked 2026-10-15 | - |
| DeepSeek-V4.1-FlashDeepSeek, released 2026-09-10 | 54.3% | MiniMax, MiniMax-M3.1-Flash model cardchecked 2026-10-25 | Dataset revision: 0808 Full release |
| Kimi K3Moonshot AI, released 2026-07-16 | 58.3% | Moonshot AI, Kimi K3 model cardchecked 2026-10-15; printed on 1 more page | Effort: max |
| MiniMax-M3.1-FlashMiniMax, released 2026-10-19 | 61.7% | MiniMax, MiniMax-M3.1-Flash model cardchecked 2026-10-25; printed on 1 more page | Dataset revision: 0808 Full release |
OSWorld 2.0 partial
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-5.6 SolOpenAI, released 2026-07-09 | 62.6% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15 | Tools: batch tool enabled |
| GPT-5.6 TerraOpenAI, released 2026-07-09 | 50.2% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15 | Tools: batch tool enabled |
| Claude Opus 5.5Anthropic, released 2026-09-22 | 81.8% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25 | Effort: max |
| Claude Fable 5.1Anthropic, released 2026-09-01 | 77.9% | Anthropic, Claude Fable 5.1 and Mythos 5.1checked 2026-10-15; printed on 2 more pages | - |
| Claude Opus 5Anthropic, released 2026-07-24 | 74% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 3 more pages | Effort: max |
| Claude Sonnet 5Anthropic, released 2026-06-30 | 42.6% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15 | Tools: batch tool enabled |
| Claude Fable 5Anthropic, released 2026-06-09 | 72.9% | Anthropic, Claude Fable 5.1 and Mythos 5.1checked 2026-10-15; printed on 1 more page | - |
| Gemini 3.8 FlashGoogle, released 2026-09-02 | 59% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15 | Tools: batch tool enabled |
| Gemini 3.7 FlashGoogle, released 2026-08-13 | 50.6% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15 | Tools: batch tool enabled |
| Qwen3.7-PlusAlibaba, released 2026-06-01 | 21.5% | Alibaba, Qwen3.8-Flash-Next model cardchecked 2026-10-15 | - |
| Qwen3.8-27BAlibaba | 48% | Alibaba, Qwen3.8-Flash-Next model cardchecked 2026-10-15 | - |
| Qwen3.8-Flash-NextAlibaba | 52.3% | Alibaba, Qwen3.8-Flash-Next model cardchecked 2026-10-15 | - |
OSWorld 2.1 partial
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-6 AstraOpenAI, released 2026-09-03 | 75.8% | Meta, Meta Muse Wonder model cardchecked 2026-10-19; printed on 1 more page | Effort: max |
| Claude Sonnet 5.5Anthropic, released 2026-09-28 | 80.1% | Anthropic, Claude Sonnet 5.5checked 2026-10-14; printed on 1 more page | - |
| Claude Opus 5.5Anthropic, released 2026-09-22 | 81.8% | Anthropic, Claude Opus 5.5checked 2026-10-14; printed on 4 more pages | - |
| Claude Fable 5.1Anthropic, released 2026-09-01 | 80.7% | Anthropic, Claude Opus 5.5checked 2026-10-14; printed on 2 more pages | - |
| Claude Opus 5Anthropic, released 2026-07-24 | 74% | Anthropic, Claude Opus 5.5checked 2026-10-14 | - |
| Claude Sonnet 5Anthropic, released 2026-06-30 | 57% | Anthropic, Claude Sonnet 5.5checked 2026-10-14; printed on 1 more page | - |
| Muse WonderMeta, released 2026-10-12 | 72.6% | Meta, Meta Muse Wonder model cardchecked 2026-10-19 | Effort: max |
| Muse Wonder FullMeta, released 2026-10-12 | 78.4% | Meta, Meta Muse Wonder model cardchecked 2026-10-19 | Effort: max |
| DeepSeek-V4.1-Flash-1017DeepSeek, released 2026-10-17 | 69.3% | Moonshot AI, Introducing Kimi K3.1 Prochecked 2026-10-24 | Effort: max |
| Kimi K3.1 ProMoonshot AI, released 2026-10-23 | 79.6% | Moonshot AI, Introducing Kimi K3.1 Prochecked 2026-10-24 | Effort: max; Harness: Kimi Agent |
OSWorld 2.1 offline subset
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-6 LunaOpenAI, released 2026-09-22 | 48.9% | Anthropic, Claude Haiku 5.5checked 2026-10-15; printed on 1 more page | - |
| GPT-6 AstraOpenAI, released 2026-09-03 | 80.9% | Google, Gemini 4 Carbonchecked 2026-10-29 | - |
| Claude Haiku 5.5Anthropic, released 2026-10-07 | 72.4% | Anthropic, Claude Haiku 5.5checked 2026-10-15; printed on 1 more page | - |
| Claude Sonnet 5.5Anthropic, released 2026-09-28 | 83.9% | Anthropic, Claude Haiku 5.5checked 2026-10-15; printed on 1 more page | - |
| Claude Opus 5.5Anthropic, released 2026-09-22 | 87.2% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19; printed on 1 more page | - |
| Claude Fable 5.1Anthropic, released 2026-09-01 | 76.8% | Google, Gemini 4 Carbonchecked 2026-10-29 | - |
| Claude Haiku 4.5Anthropic, released 2025-10-15 | 15.7% | Anthropic, Claude Haiku 5.5checked 2026-10-15; printed on 1 more page | - |
| Gemini 4 CarbonGoogle, released 2026-10-28 | 81.5% | Google, Gemini 4 Carbonchecked 2026-10-29 | Harness: Gemini CUA harness |
| Gemini 4 ArgonGoogle, released 2026-09-30 | 74.7% | Google, Gemini 4 Carbonchecked 2026-10-29 | Harness: Gemini CUA harness |
OSWorld 2.1 binary
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-6 AstraOpenAI, released 2026-09-03 | 34.3% | Meta, Meta Muse Wonder model cardchecked 2026-10-19 | Effort: max |
| Claude Haiku 5.5Anthropic, released 2026-10-07 | 37.1% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
| Claude Sonnet 5.5Anthropic, released 2026-09-28 | 43.5% | Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19; printed on 1 more page | Effort: max; Context: Compaction above 100k tokens |
| Claude Opus 5.5Anthropic, released 2026-09-22 | 53.2% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19; printed on 1 more page | - |
| Claude Fable 5.1Anthropic, released 2026-09-01 | 43.1% | Meta, Meta Muse Wonder model cardchecked 2026-10-19 | Effort: max |
| Claude Sonnet 5Anthropic, released 2026-06-30 | 25.6% | Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19 | Effort: max; Context: Compaction above 100k tokens |
| Muse WonderMeta, released 2026-10-12 | 34.3% | Meta, Meta Muse Wonder model cardchecked 2026-10-19 | Effort: max |
| Muse Wonder FullMeta, released 2026-10-12 | 39.4% | Meta, Meta Muse Wonder model cardchecked 2026-10-19 | Effort: max |
OSWorld 2.0 strict
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Claude Opus 5.5Anthropic, released 2026-09-22 | 48.7% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25 | Effort: max |
| Claude Fable 5.1Anthropic, released 2026-09-01 | 41.7% | Anthropic, Claude Fable 5.1 and Mythos 5.1checked 2026-10-15; printed on 2 more pages | - |
| Claude Opus 5Anthropic, released 2026-07-24 | 37.2% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 2 more pages | Effort: max |
| Claude Fable 5Anthropic, released 2026-06-09 | 36.1% | Anthropic, Claude Fable 5.1 and Mythos 5.1checked 2026-10-15; printed on 1 more page | - |
| Qwen3.7-PlusAlibaba, released 2026-06-01 | 2.8% | Alibaba, Qwen3.8-Flash-Next model cardchecked 2026-10-15 | - |
| Qwen3.8-27BAlibaba | 19.4% | Alibaba, Qwen3.8-Flash-Next model cardchecked 2026-10-15 | - |
| Qwen3.8-Flash-NextAlibaba | 19.4% | Alibaba, Qwen3.8-Flash-Next model cardchecked 2026-10-15 | - |
OSWorld 2.0 offline partial
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-6 SolOpenAI, released 2026-09-22 | 60.5% | OpenAI, Introducing GPT-6 Sol and GPT-6 Lunachecked 2026-10-19 | Effort: xhigh |
| GPT-6 AstraOpenAI, released 2026-09-03 | 72.6% | OpenAI, GPT-6 Astrachecked 2026-10-14; printed on 1 more page | - |
| GPT-5.6 SolOpenAI, released 2026-07-09 | 65.7% | OpenAI, GPT-6 Astrachecked 2026-10-14 | - |
| Claude Opus 5Anthropic, released 2026-07-24 | 70.2% | OpenAI, GPT-6 Astrachecked 2026-10-14; printed on 1 more page | - |
| Gemini 4 ArgonGoogle, released 2026-09-30 | 69.2% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.