# Vals SAGE (SAGE)

> Student Assessment with Generative Evaluation

Canonical page: https://benchlm.ai/benchmarks/sage

- Category: [Multimodal & Grounded](/multimodal-grounded)
- Last updated: October 8, 2026

Evaluation results published by [Vals AI](https://www.vals.ai/benchmarks/sage). Published leaderboard aggregates and evaluation settings only. Dataset contents are not included.

## About SAGE

- Year: 2026
- Tasks: Student assessment with generative evaluation
- Format: Accuracy score
- Difficulty: Education assessment reasoning
- Paper: [SAGE](https://www.vals.ai/benchmarks/sage)

BenchLM mirrors the public Vals AI SAGE leaderboard as display-only external evidence. The captured snapshot preserves overall scores, task-level scores where Vals publishes them, uncertainty, latency, and cost-per-test metadata. It is excluded from BenchLM weighted rankings.

SAGE is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (91 models)

| Rank | Model | Configuration | Creator | Score |
|------|-------|---------------|---------|-------|
| 1 | [Claude Opus 4.7](/models/claude-opus-4-7) | — | Anthropic | 56.10% |
| 2 | [Gemma 4 31b It](https://www.vals.ai/models/google_gemma-4-31b-it) | high reasoning | Google | 55.03% |
| 3 | [Claude Opus 4.8](/models/claude-opus-4-8) | — | Anthropic | 54.79% |
| 4 | [Gemini 4 Argon](/models/gemini-4-argon) | high reasoning | Google | 53.65% |
| 5 | [Kimi K3](/models/kimi-k3) | max reasoning | Moonshot AI | 52.78% |
| 6 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | max reasoning | OpenAI | 52.56% |
| 7 | [Qwen3.8-27B](/models/qwen3-8-27b) | xhigh reasoning | Alibaba | 52.40% |
| 8 | [Claude Opus 4.5 Thinking](/models/claude-opus-4-5-thinking) | — | Anthropic | 52.09% |
| 9 | [Claude Fable 5](/models/claude-fable) | — | Anthropic | 51.89% |
| 10 | [Gemini 3 Flash Preview](https://www.vals.ai/models/google_gemini-3-flash-preview) | high reasoning | Google | 51.85% |
| 11 | [Claude Sonnet 5.5](/models/claude-sonnet-5-5) | — | Anthropic | 51.77% |
| 12 | [Claude Opus 4.6 (Adaptive)](/models/claude-opus-4-6-thinking) | — | Anthropic | 51.58% |
| 13 | [GPT-5.5](/models/gpt-5-5) | xhigh reasoning | OpenAI | 51.53% |
| 14 | [Qwen3.8 Max](/models/qwen3-8-max) | — | Alibaba | 51.25% |
| 15 | [GPT-5.4 mini](/models/gpt-5-4-mini) | xhigh reasoning | OpenAI | 50.81% |
| 16 | [MiniMax M3](/models/minimax-m3) | — | MiniMax | 50.57% |
| 17 | [Kimi K2.6](/models/kimi-2-6) | — | Moonshot AI | 50.22% |
| 18 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | high reasoning | Google | 49.88% |
| 19 | [Kimi K2.5 Thinking](https://www.vals.ai/models/kimi_kimi-k2.5-thinking) | — | Moonshot AI | 49.87% |
| 20 | [Gemini 3.1 Flash Lite Preview](https://www.vals.ai/models/google_gemini-3.1-flash-lite-preview) | high reasoning | Google | 49.54% |
| 21 | [Claude Opus 5](/models/claude-opus-5) | — | Anthropic | 49.43% |
| 22 | [GPT-5.2](/models/gpt-5-2) | xhigh reasoning | OpenAI | 49.27% |
| 23 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | high reasoning | Google | 49.23% |
| 24 | [Claude Sonnet 5](/models/claude-sonnet-5) | — | Anthropic | 48.92% |
| 25 | [Gemini 3.6 Flash](/models/gemini-3-6-flash) | high reasoning | Google | 48.69% |
| 26 | [Gemini 3.1 Pro Preview](https://www.vals.ai/models/google_gemini-3.1-pro-preview) | high reasoning | Google | 48.68% |
| 27 | [Claude Fable 5.1](/models/claude-fable-5-1) | — | Anthropic | 48.53% |
| 28 | [GPT-6 Luna](/models/gpt-6-luna) | max reasoning | OpenAI | 48.09% |
| 29 | [DeepSeek V4.1 Flash 1017](/models/deepseek-v4-1-flash) | high reasoning | DeepSeek | 47.88% |
| 30 | [Muse Spark 1.2](/models/muse-spark-1-2) | xhigh reasoning | Meta | 47.66% |
| 31 | [Gemini 3 Pro Preview](https://www.vals.ai/models/google_gemini-3-pro-preview) | high reasoning | Google | 47.62% |
| 32 | [Gemini 3.5 Flash-Lite](/models/gemini-3-5-flash-lite) | high reasoning | Google | 47.30% |
| 33 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | max reasoning | OpenAI | 47.00% |
| 34 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | — | Anthropic | 46.58% |
| 35 | [GPT-6.1 Sol](/models/gpt-6-1-sol) | max reasoning | OpenAI | 46.54% |
| 36 | [GPT-6 Astra](/models/gpt-6-astra) | max reasoning | OpenAI | 46.37% |
| 37 | [Claude Opus 5.5](/models/claude-opus-5-5) | — | Anthropic | 45.83% |
| 38 | [Claude Haiku 5.5](/models/claude-haiku-5-5) | — | Anthropic | 45.78% |
| 39 | [Muse Spark 1.1](/models/muse-spark-1-1) | xhigh reasoning | Meta | 45.58% |
| 40 | [Qwen3.6-27B](/models/qwen3-6-27b) | — | Alibaba | 45.58% |
| 41 | [MiMo-V2.6-Pro](/models/mimo-v2-6-pro) | — | Xiaomi | 45.05% |
| 42 | [Claude Opus 4.5](/models/claude-opus-4-5) | — | Anthropic | 45.00% |
| 43 | [Qwen3.6 Plus](/models/qwen3-6-plus) | — | Alibaba | 44.86% |
| 44 | [GPT-6 Sol](/models/gpt-6-sol) | max reasoning | OpenAI | 44.79% |
| 45 | [Gemini 2.5 Flash](/models/gemini-2-5-flash) | — | Google | 44.76% |
| 46 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | max reasoning | OpenAI | 44.22% |
| 47 | [Gemini 2.5 Flash Preview 09 2025](https://www.vals.ai/models/google_gemini-2.5-flash-preview-09-2025) | — | Google | 44.11% |
| 48 | [GPT-5](https://www.vals.ai/models/openai_gpt-5-2025-08-07) | high reasoning | OpenAI | 43.68% |
| 49 | [MiMo-V2.6-Flash](/models/mimo-v2-6-flash) | — | Xiaomi | 43.53% |
| 50 | [GPT-5.4](/models/gpt-5-4) | xhigh reasoning | OpenAI | 43.31% |
| 51 | [MiMo-V2.5](/models/mimo-v2-5) | — | Xiaomi | 43.27% |
| 52 | [GPT-5.1](/models/gpt-5-1) | high reasoning | OpenAI | 43.23% |
| 53 | [GPT-5 mini](/models/gpt-5-mini) | high reasoning | OpenAI | 42.99% |
| 54 | [Qwen3.5 Flash](/models/qwen3-5-flash) | — | Alibaba | 42.49% |
| 55 | [Qwen3 VL Plus](https://www.vals.ai/models/alibaba_qwen3-vl-plus-2025-09-23) | — | Alibaba | 42.10% |
| 56 | [Gemini 2.5 Pro](/models/gemini-2-5-pro) | — | Google | 41.92% |
| 57 | [o3](/models/o3) | high reasoning | OpenAI | 41.77% |
| 58 | [o4-mini](/models/o4-mini) | high reasoning | OpenAI | 41.06% |
| 59 | [Grok 4.7](/models/grok-4-7) | xhigh reasoning | xAI | 40.79% |
| 60 | [Gemini 2.5 Flash Thinking](https://www.vals.ai/models/google_gemini-2.5-flash-thinking) | — | Google | 39.60% |
| 61 | [Gemini 2.5 Flash Preview 09 2025 Thinking](https://www.vals.ai/models/google_gemini-2.5-flash-preview-09-2025-thinking) | — | Google | 39.38% |
| 62 | [Qwen3.7 Plus](/models/qwen3-7-plus) | — | Alibaba | 39.25% |
| 63 | [Grok 4.20 0309 Reasoning](https://www.vals.ai/models/grok_grok-4.20-0309-reasoning) | — | xAI | 38.24% |
| 64 | [GPT-5.4 nano](/models/gpt-5-4-nano) | high reasoning | OpenAI | 38.08% |
| 65 | [Mistral Medium 3.5](https://www.vals.ai/models/mistralai_mistral-medium-3.5) | high reasoning | Mistral AI | 37.61% |
| 66 | [Inkling](/models/inkling) | 0.99 reasoning | Thinking Machines Lab | 36.55% |
| 67 | [Claude Sonnet 4.5 Thinking](/models/claude-sonnet-4-5-thinking) | — | Anthropic | 36.06% |
| 68 | [Llama4 Maverick Instruct Basic](https://www.vals.ai/models/fireworks_llama4-maverick-instruct-basic) | — | Fireworks AI | 35.70% |
| 69 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | high reasoning | Google | 35.06% |
| 70 | [Claude Sonnet 4](https://www.vals.ai/models/anthropic_claude-sonnet-4-20250514) | — | Anthropic | 35.00% |
| 71 | [Grok 4.5](/models/grok-4-5) | high reasoning | xAI | 34.97% |
| 72 | [Meta Llama Llama 4 Scout 17B 16E Instruct](https://www.vals.ai/models/together_meta-llama/Llama-4-Scout-17B-16E-Instruct) | — | Together AI | 34.83% |
| 73 | [Claude Sonnet 4.5](/models/claude-sonnet-4-5) | — | Anthropic | 32.88% |
| 74 | [Inkling-Small](/models/inkling-small) | 0.99 reasoning | Thinking Machines Lab | 32.74% |
| 75 | [Claude Sonnet 4 20250514 Thinking](https://www.vals.ai/models/anthropic_claude-sonnet-4-20250514-thinking) | — | Anthropic | 32.01% |
| 76 | [Claude Haiku 4.5 Thinking](/models/claude-haiku-4-5-thinking) | — | Anthropic | 31.82% |
| 77 | [Claude Opus 4.1](https://www.vals.ai/models/anthropic_claude-opus-4-1-20250805) | — | Anthropic | 31.38% |
| 78 | [Grok 4.1 Fast (Reasoning)](/models/grok-4-1-fast-reasoning) | — | xAI | 31.33% |
| 79 | [Gemini 2.5 Flash Lite Preview 09 2025 Thinking](https://www.vals.ai/models/google_gemini-2.5-flash-lite-preview-09-2025-thinking) | — | Google | 30.78% |
| 80 | [Qwen3.5 Plus Thinking](https://www.vals.ai/models/alibaba_qwen3.5-plus-thinking) | — | Alibaba | 30.40% |
| 81 | [Claude Opus 4.1 20250805 Thinking](https://www.vals.ai/models/anthropic_claude-opus-4-1-20250805-thinking) | — | Anthropic | 30.39% |
| 82 | [GPT-5 nano](/models/gpt-5-nano) | high reasoning | OpenAI | 30.38% |
| 83 | [Grok 4 Fast (Reasoning)](/models/grok-4-fast-reasoning) | — | xAI | 29.76% |
| 84 | [Grok 4.6](/models/grok-4-6) | high reasoning | xAI | 28.90% |
| 85 | [Gemini 2.5 Flash Lite Preview 09 2025](https://www.vals.ai/models/google_gemini-2.5-flash-lite-preview-09-2025) | — | Google | 27.97% |
| 86 | [Gemini 2.5 Flash-Lite](/models/gemini-2-5-flash-lite) | — | Google | 27.53% |
| 87 | [Grok 4 0709](https://www.vals.ai/models/grok_grok-4-0709) | — | xAI | 25.10% |
| 88 | [Mistral Large 2512](https://www.vals.ai/models/mistralai_mistral-large-2512) | — | Mistral AI | 24.59% |
| 89 | [Grok 4.3](/models/grok-4-3) | — | xAI | 19.74% |
| 90 | [Grok 4 Fast Non Reasoning](https://www.vals.ai/models/grok_grok-4-fast-non-reasoning) | — | xAI | 15.99% |
| 91 | [Grok 4.1 Fast Non Reasoning](https://www.vals.ai/models/grok_grok-4-1-fast-non-reasoning) | — | xAI | 11.91% |

## FAQ

### What does SAGE measure?

Student Assessment with Generative Evaluation

### Which model leads the published SAGE snapshot?

Claude Opus 4.7 currently leads the published SAGE snapshot with a score of 56.10%.

### How many models are evaluated on SAGE?

The October 8, 2026 contains 91 AI models.

### Does SAGE affect BenchLM's overall score?

Not directly. SAGE is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.