Instruction Following / Benchmark profile
A benchmark of 541 prompts built from 25 verifiable instruction types. It tests whether a model follows checkable constraints such as keyword, length, casing, and response-format requirements.
Observed results
20 ranked models · higher is better · labels show rank and score
| Model | Rank | Model ID | Provider | Score | Unit |
|---|---|---|---|---|---|
| Qwen3.5 27B | #1 | qwen3-5-27b | Alibaba Cloud | 95 | percent |
| Agents-A1 | #2 | agents-a1 | InternScience | 94.82 | percent |
| Qwen3.7-Plus | #3 | qwen-3-7-plus | Alibaba Cloud | 94.6 | percent |
| Qwen3.6 Plus | #4 | qwen3-6-plus | Alibaba Cloud | 94.3 | percent |
| Qwen3.7-Max | #5 | qwen-3-7-max | Alibaba Cloud | 94.3 | percent |
| Kimi K2.5 | #6 | kimi-k2-5 | Moonshot AI | 93.9 | percent |
| o3-mini | #7 | o3-mini | OpenAI | 93.9 | percent |
| Qwen3.5-122B-A10B | #8 | qwen3-5-122b-a10b | Alibaba Cloud | 93.4 | percent |
| GLM-5 | #9 | glm-5 | Z.AI | 92.6 | percent |
| Qwen3.5 397B A17B | #10 | qwen3-5-397b | Alibaba Cloud | 92.6 | percent |
| o1 | #11 | o1 | OpenAI | 92.2 | percent |
| Qwen3.5-35B-A3B | #12 | qwen3-5-35b-a3b | Alibaba Cloud | 91.9 | percent |
| LFM2.5-8B-A1B | #13 | lfm2-5-8b-a1b | LiquidAI | 91.84 | percent |
| Claude Opus 4.5 | #14 | claude-opus-4-5 | Anthropic | 90.9 | percent |
| GPT-4.1 mini | #15 | gpt-4-1-mini | OpenAI | 88.5 | percent |
| GPT-4.1 | #16 | gpt-4-1 | OpenAI | 87.4 | percent |
| DeepSeek V3 | #17 | deepseek-v3 | DeepSeek | 86.1 | percent |
| ZAYA1-8B | #18 | zaya1-8b | Zyphra | 85.58 | percent |
| GPT-4.1 nano | #19 | gpt-4-1-nano | OpenAI | 83.2 | percent |
| MiniCPM5-1B | #20 | minicpm5-1b | OpenBMB | 80.41 | percent |
| Mellum2-12B-A2.5B-Thinking | #21 | mellum2-12b-a2-5b-thinking | JetBrains | 76.5 | percent |
| Mellum2-12B-A2.5B-Instruct | #22 | mellum2-12b-a2-5b-instruct | JetBrains | 75.8 | percent |
| LFM2.5-230M | #23 | lfm2-5-230m | LiquidAI | 71.71 | percent |
| LFM2.5-VL-450M | #24 | lfm2-5-vl-450m | LiquidAI | 61.16 | percent |
One best compatible score per canonical product · higher is better
| Rank | Model | Weights | Evidence | Score |
|---|---|---|---|---|
| 01 | Qwen3.5 27BAlibaba Cloud | Open weights | Source carrier | 95% |
| 02 | Agents-A1InternScience | Open weights | Source carrier | 94.8% |
| 03 | Qwen3.7-PlusAlibaba Cloud | Closed weights | Source carrier | 94.6% |
| 04 | Qwen3.6 PlusAlibaba Cloud | Closed weights | Source carrier | 94.3% |
| 05 | Qwen3.7-MaxAlibaba Cloud | Closed weights | Source carrier | 94.3% |
| 06 | Kimi K2.5Moonshot AI | Open weights | Source carrier | 93.9% |
| 07 | o3-miniOpenAI | Closed weights | Source carrier | 93.9% |
| 08 | Qwen3.5-122B-A10BAlibaba Cloud | Open weights | Source carrier | 93.4% |
| 09 | GLM-5Z.AI | Open weights | Source carrier | 92.6% |
| 10 | Qwen3.5 397B A17BAlibaba Cloud | Open weights | Source carrier | 92.6% |
Each distinct published measurement is retained. Same-snapshot canonical and configuration projections appear once.
| Published model / configuration | Score | Evidence & protocol | Source & dates |
|---|---|---|---|
Exact BenchLM registry variant Qwen3.5-27B; bulk export does not retain a complete upstream harness configuration.Qwen3.5 27BExact identitySource label without a registered configuration ID Canonical product: qwen3-5-27b | 95% | Source carriersource-checkedVersion & system2023 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Agents-A1; bulk export does not retain a complete upstream harness configuration.Agents-A1Exact identitySource label without a registered configuration ID Canonical product: agents-a1 | 94.8% | Source carriersource-checkedVersion & system2023 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Qwen3.7 Plus; bulk export does not retain a complete upstream harness configuration.Qwen3.7-PlusExact identitySource label without a registered configuration ID Canonical product: qwen-3-7-plus | 94.6% | Source carriersource-checkedVersion & system2023 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Qwen3.6 Plus; bulk export does not retain a complete upstream harness configuration.Qwen3.6 PlusExact identitySource label without a registered configuration ID Canonical product: qwen3-6-plus | 94.3% | Source carriersource-checkedVersion & system2023 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Qwen3.7 Max; bulk export does not retain a complete upstream harness configuration.Qwen3.7-MaxExact identitySource label without a registered configuration ID Canonical product: qwen-3-7-max | 94.3% | Source carriersource-checkedVersion & system2023 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Kimi K2.5; bulk export does not retain a complete upstream harness configuration.Kimi K2.5Exact identitySource label without a registered configuration ID Canonical product: kimi-k2-5 | 93.9% | Source carriersource-checkedVersion & system2023 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant o3-mini; bulk export does not retain a complete upstream harness configuration.o3-miniExact identitySource label without a registered configuration ID Canonical product: o3-mini | 93.9% | Source carriersource-checkedVersion & system2023 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Qwen3.5-122B-A10B; bulk export does not retain a complete upstream harness configuration.Qwen3.5-122B-A10BExact identitySource label without a registered configuration ID Canonical product: qwen3-5-122b-a10b | 93.4% | Source carriersource-checkedVersion & system2023 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant GLM-5; bulk export does not retain a complete upstream harness configuration.GLM-5Exact identitySource label without a registered configuration ID Canonical product: glm-5 | 92.6% | Source carriersource-checkedVersion & system2023 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Qwen3.5 397B; bulk export does not retain a complete upstream harness configuration.Qwen3.5 397B A17BExact identitySource label without a registered configuration ID Canonical product: qwen3-5-397b | 92.6% | Source carriersource-checkedVersion & system2023 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
From result to context
A benchmark of 541 prompts built from 25 verifiable instruction types. It tests whether a model follows checkable constraints such as keyword, length, casing, and response-format requirements.
The summary shows one best compatible source result per canonical product. Versions, effort settings and execution systems remain attached to the underlying records.
Scores from different versions or harnesses may not be interchangeable. The published source rows preserve those distinctions and their original units. A source result is not automatically an input to a current index.
This catalogue entry preserves available source evidence. Current indices admit only their specifically reviewed tracks and configurations.
Contamination risk: Unknown. Lifecycle: Active.
