BlogLLM Benchmark Scores as Published, Side by Side
IFEval: Results as Published
Every result the catalog holds for IFEval: 29 models from 9 developers, read from 7 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 Instruction FollowingMaintainer Google (Jeffrey Zhou and co-authors)Introduced 2023Data version 2026103001, as of 2026-10-30
IFEval tests instruction following with about 500 prompts that carry verifiable instructions of 25 types, such as a word limit or a required format, so compliance can be checked by code. Publishers usually print prompt-level strict accuracy: the share of prompts where every instruction was followed.
IFEval
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-5.2OpenAI, released 2025-12-11 | 94.8% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| GPT-5 miniOpenAI, released 2025-08-07 | 93.9% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| gpt-oss-120bOpenAI, released 2025-08-05 | 88.9% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| GPT-4oOpenAI, released 2024-05-13 | 87.2% | Mistral AI, Medium is the new largechecked 2026-10-15 | - |
| Claude Opus 4.5Anthropic, released 2025-11-24 | 90.9% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| Claude 3.7 SonnetAnthropic, released 2025-02-24 | 93.2% | Anthropic, Claude 3.7 Sonnetchecked 2026-10-15; printed on 1 more page | Effort: extended thinking, 64K |
| 90.8% | Anthropic, Claude 3.7 Sonnetchecked 2026-10-15 | Effort: none | |
| Claude 3.5 SonnetAnthropic, released 2024-10-22 | 90.2% | Anthropic, Claude 3.7 Sonnetchecked 2026-10-15 | - |
| Gemini 3 ProGoogle, released 2025-11-18 | 93.5% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| Llama 4 MaverickMeta, released 2025-04-05 | 88.9% | Mistral AI, Medium is the new largechecked 2026-10-15 | - |
| DeepSeek-V3-0324DeepSeek, released 2025-03-24 | 89.1% | Mistral AI, Medium is the new largechecked 2026-10-15 | - |
| DeepSeek-R1DeepSeek, released 2025-01-20 | 83.3% | Anthropic, Claude 3.7 Sonnetchecked 2026-10-15 | Effort: extended thinking, 32K |
| Qwen3.5-122B-A10BAlibaba, released 2026-02-24 | 93.4% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| Qwen3.5-27BAlibaba, released 2026-02-24 | 95% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| Qwen3.5-35B-A3BAlibaba, released 2026-02-24 | 91.9% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| Qwen3.5-397B-A17BAlibaba, released 2026-02-16 | 92.6% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| Qwen3-30B-A3B-Instruct-2507Alibaba, released 2025-07-30 | 91% | AI21 Labs, AI21 Jamba2 Mini model cardchecked 2026-10-15 | - |
| Qwen3-235B-A22BAlibaba, released 2025-04-29 | 87.8% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| Qwen3-Max-ThinkingAlibaba | 93.4% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| Kimi K2.5Moonshot AI, released 2026-01-27 | 93.9% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| Ministral 3 14B Instruct 2512Mistral AI, released 2025-12-02 | 73% | AI21 Labs, AI21 Jamba2 Mini model cardchecked 2026-10-15 | - |
| Mistral Medium 3Mistral AI, released 2025-05-07 | 89.4% | Mistral AI, Medium is the new largechecked 2026-10-15 | - |
| Nemotron 3 Nano 30B-A3BNVIDIA, released 2025-12-15 | 88% | AI21 Labs, AI21 Jamba2 Mini model cardchecked 2026-10-15 | Effort: non-thinking |
IFEval (prompt-level strict)
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-4.1OpenAI, released 2025-04-14 | 88% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| o1-miniOpenAI, released 2024-09-12 | 84.8% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | - |
| GPT-4oOpenAI, released 2024-05-13 | 84.3% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | - |
| Claude Opus 4Anthropic, released 2025-05-22 | 87.4% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Effort: no extended thinking |
| Claude Sonnet 4Anthropic, released 2025-05-22 | 87.6% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Effort: no extended thinking |
| Claude 3.5 SonnetAnthropic, released 2024-10-22 | 86.5% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | - |
| Gemini 2.5 FlashGoogle, released 2025-05-20 | 84.3% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| DeepSeek-V3-0324DeepSeek, released 2025-03-24 | 81.1% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| DeepSeek-R1DeepSeek, released 2025-01-20 | 83.3% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | - |
| DeepSeek-V3DeepSeek, released 2024-12-26 | 86.1% | DeepSeek, DeepSeek-R1 model cardchecked 2026-10-15 | - |
| Qwen3-235B-A22BAlibaba, released 2025-04-29 | 83.2% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Effort: non-thinking |
| Kimi K2 InstructMoonshot AI, released 2025-07-11 | 89.8% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
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.