Skip to main content

Vals SAGE (SAGE)

Student Assessment with Generative Evaluation

Data verified

How BenchLM shows SAGE

BenchLM mirrors the public Vals AI SAGE leaderboard captured from https://www.vals.ai/benchmarks/sage and updated by Vals on July 9, 2026. The snapshot preserves overall scores, uncertainty, latency, cost-per-test metadata, and task-level scores where Vals publishes them.

SAGE is display only on BenchLM. Vals proprietary or Vals-hosted aggregate views are useful context, but BenchLM does not use them as weighted ranking inputs or as a replacement for benchmark-native source records.

66 Vals rows5 task viewsprivate datasetTasks: Overall, Partial Differential Equations, Linear Algebra, Automata, CalculusDisplay only

SAGE score on SAGE — July 9, 2026

BenchLM mirrors the published sage score view for SAGE. Claude Opus 4.7 leads the public snapshot at 56.10% , followed by Gemma 4 31b It (55.03%) and Claude Opus 4.8 (54.79%). BenchLM does not use these results to rank models overall.

66 modelsExternal benchmark mirrorsCurrentDisplay onlyUpdated July 9, 2026

The published SAGE snapshot is tightly clustered at the top: Claude Opus 4.7 sits at 56.10%, while the third row is only 1.32 points behind. The broader top-10 spread is 5.29 points, so many of the published scores sit in a relatively narrow band.

66 models have been evaluated on SAGE. The benchmark falls in the External benchmark mirrors category. BenchLM tracks this category separately from its weighted global scoring system, so these results are best compared on the dedicated Korean benchmark views. SAGE is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About SAGE

Year

2026

Tasks

Student assessment with generative evaluation

Format

Accuracy score

Difficulty

Education assessment reasoning

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.

BenchLM freshness & provenance

Version

SAGE 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

SAGE score table (66 models)

1
Claude Opus 4.7anthropic/claude-opus-4-7
56.10%
2
Gemma 4 31b Itgoogle/gemma-4-31b-it
55.03%
3
Claude Opus 4.8anthropic/claude-opus-4-8
54.79%
4
GPT-5.6 Solopenai/gpt-5.6-sol
52.56%
5
Claude Opus 4.5 20251101 Thinkinganthropic/claude-opus-4-5-20251101-thinking
52.09%
6
Claude Fable 5anthropic/claude-fable-5
51.89%
7
Gemini 3 Flash Previewgoogle/gemini-3-flash-preview
51.85%
8
Claude Opus 4.6 Thinkinganthropic/claude-opus-4-6-thinking
51.58%
9
GPT-5.5openai/gpt-5.5
51.53%
10
GPT-5.4 Miniopenai/gpt-5.4-mini-2026-03-17
50.81%
11
MiniMax M3minimax/MiniMax-M3
50.57%
12
Kimi K2.6kimi/kimi-k2.6
50.22%
13
Gemini 3.5 Flashgoogle/gemini-3.5-flash
49.88%
14
Kimi K2.5 Thinkingkimi/kimi-k2.5-thinking
49.87%
15
Gemini 3.1 Flash Lite Previewgoogle/gemini-3.1-flash-lite-preview
49.54%
16
GPT-5.2openai/gpt-5.2-2025-12-11
49.27%
17
Claude Sonnet 5anthropic/claude-sonnet-5
48.92%
18
Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview
48.68%
19
Gemini 3 Pro Previewgoogle/gemini-3-pro-preview
47.62%
20
GPT-5.6 Terraopenai/gpt-5.6-terra
47.00%
21
Claude Sonnet 4.6anthropic/claude-sonnet-4-6
46.58%
22
Muse Spark 1 1meta/muse_spark_1_1
45.58%
23
Qwen3.6 27balibaba/qwen3.6-27b
45.58%
24
Claude Opus 4.5anthropic/claude-opus-4-5-20251101
45.00%
25
Qwen3.6 Plusalibaba/qwen3.6-plus
44.86%
26
Gemini 2.5 Flashgoogle/gemini-2.5-flash
44.76%
27
GPT-5.6 Lunaopenai/gpt-5.6-luna
44.22%
28
Gemini 2.5 Flash Preview 09 2025google/gemini-2.5-flash-preview-09-2025
44.11%
29
GPT-5openai/gpt-5-2025-08-07
43.68%
30
GPT-5.4openai/gpt-5.4-2026-03-05
43.31%
31
Mimo V2.5xiaomi/mimo-v2.5
43.27%
32
GPT-5.1openai/gpt-5.1-2025-11-13
43.23%
33
GPT-5 Miniopenai/gpt-5-mini-2025-08-07
42.99%
34
Qwen3.5 Flashalibaba/qwen3.5-flash
42.49%
35
Qwen3 VL Plusalibaba/qwen3-vl-plus-2025-09-23
42.10%
36
Gemini 2.5 Progoogle/gemini-2.5-pro
41.92%
37
O3openai/o3-2025-04-16
41.77%
38
O4 Miniopenai/o4-mini-2025-04-16
41.06%
39
Gemini 2.5 Flash Thinkinggoogle/gemini-2.5-flash-thinking
39.60%
40
Gemini 2.5 Flash Preview 09 2025 Thinkinggoogle/gemini-2.5-flash-preview-09-2025-thinking
39.38%
41
Qwen3.7 Plusalibaba/qwen3.7-plus
39.25%
42
Grok 4.20 0309 Reasoninggrok/grok-4.20-0309-reasoning
38.24%
43
GPT-5.4 Nanoopenai/gpt-5.4-nano-2026-03-17
38.08%
44
Mistral Medium 3.5mistralai/mistral-medium-3.5
37.61%
45
Claude Sonnet 4.5 20250929 Thinkinganthropic/claude-sonnet-4-5-20250929-thinking
36.06%
46
Llama4 Maverick Instruct Basicfireworks/llama4-maverick-instruct-basic
35.70%
47
Claude Sonnet 4anthropic/claude-sonnet-4-20250514
35.00%
48
Meta Llama Llama 4 Scout 17B 16E Instructtogether/meta-llama/Llama-4-Scout-17B-16E-Instruct
34.83%
49
Claude Sonnet 4.5anthropic/claude-sonnet-4-5-20250929
32.88%
50
Claude Sonnet 4 20250514 Thinkinganthropic/claude-sonnet-4-20250514-thinking
32.01%
51
Claude Haiku 4.5 20251001 Thinkinganthropic/claude-haiku-4-5-20251001-thinking
31.82%
52
Claude Opus 4.1anthropic/claude-opus-4-1-20250805
31.38%
53
Grok 4.1 Fast Reasoninggrok/grok-4-1-fast-reasoning
31.33%
54
Gemini 2.5 Flash Lite Preview 09 2025 Thinkinggoogle/gemini-2.5-flash-lite-preview-09-2025-thinking
30.78%
55
Qwen3.5 Plus Thinkingalibaba/qwen3.5-plus-thinking
30.40%
56
Claude Opus 4.1 20250805 Thinkinganthropic/claude-opus-4-1-20250805-thinking
30.39%
57
GPT-5 Nanoopenai/gpt-5-nano-2025-08-07
30.38%
58
Grok 4 Fast Reasoninggrok/grok-4-fast-reasoning
29.76%
59
Gemini 2.5 Flash Lite Preview 09 2025google/gemini-2.5-flash-lite-preview-09-2025
27.97%
60
Gemini 2.5 Flash Litegoogle/gemini-2.5-flash-lite
27.53%
61
Grok 4 0709grok/grok-4-0709
25.10%
62
Mistral Large 2512mistralai/mistral-large-2512
24.59%
63
Grok 4.3grok/grok-4.3
19.74%
64
Grok 4 Fast Non Reasoninggrok/grok-4-fast-non-reasoning
15.99%
65
Grok 4.1 Fast Non Reasoninggrok/grok-4-1-fast-non-reasoning
11.91%
66
Command A Plus 05 2026cohere/command-a-plus-05-2026
10.49%

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 56.10% sage score. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on SAGE?

66 AI models are included in BenchLM's mirrored SAGE snapshot, based on the public leaderboard captured on July 9, 2026.

Last updated: July 9, 2026 · mirrored from the public benchmark leaderboard

The AI models change fast. We track them for you.

For engineers, researchers, and the plain curious — a weekly brief on new models, ranking shifts, and pricing changes.

Free. No spam. Unsubscribe anytime.