Skip to main content
Radar

Every change to the models you run, with its source and its date. Releases, price changes, retirements, API changes, and incidents.Every change to the models you run, with its source.

Follow model changes
BenchLM

GPQA Diamond (GPQA-D)

A display-only GPQA Diamond reference from provider comparison charts.

Data verified 33 confirmed releases in the last 30 daysSee provider release alerts

Benchmark score on GPQA-D — September 15, 2026

We mirror the published score view for GPQA-D. GPT-6 Astra leads the public snapshot at 96.0%, followed by Sakana Fugu-Ultra (95.5%) and Sakana Fugu (95.5%). We do not use these results to rank models overall.

1Proprietary

GPT-6 Astra

OpenAI

96.0%
Overall 82.94Context 1.05M
2Proprietary

Sakana Fugu-Ultra

Sakana AI

95.5%
Context 1M
3Proprietary

Sakana Fugu

Sakana AI

95.5%
Context 1M

67 modelsKnowledgeCurrentDisplay onlyUpdated September 15, 2026

Benchmark score table (67 models)

Score
1
GPT-6 AstraOpenAI · Closed
96.0%
2
Sakana Fugu-UltraSakana AI · Closed
95.5%
3
Sakana FuguSakana AI · Closed
95.5%
4
GPT-5.6 SolOpenAI · Closed
94.6%
5
Gemini 3.1 ProGoogle · Closed
94.3%
6
Claude Opus 4.7 (Adaptive)Anthropic · Closed
94.2%
7
Claude Opus 4.8Anthropic · Closed
93.6%
8
GPT-5.5OpenAI · Closed
93.6%
9
Kimi K3Moonshot AI · Closed
93.5%
10
GPT-5.6 TerraOpenAI · Closed
92.9%
11
GPT-5.4OpenAI · Closed
92.8%
12
Ornith-1.5-397BOrnith AI · Open weight
92.8%
13
Gemini 3.5 FlashGoogle · Closed
92.7%
14
Qwen3.8 MaxAlibaba · Open weight
92.6%
15
Qwen3.7 MaxAlibaba · Closed
92.4%
16
GPT-5.6 LunaOpenAI · Closed
92.3%
17
Hy4 previewTencent · Open weight
92.3%
18
Qwen3.8-Flash-NextAlibaba · Open weight
91.7%
19
GLM-5.2Z.AI · Open weight
91.2%
20
DeepSeek V4.1 FlashDeepSeek · Open weight
90.9%
21
Kimi K2.6Moonshot AI · Open weight
90.5%
22
Qwen3.7 PlusAlibaba · Closed
90.3%
23
DeepSeek V4 Pro 0813DeepSeek · Closed
90.1%
24
Interfaze BetaInterfaze · Closed
89.9%
25
Inkling-SmallThinking Machines Lab · Open weight
89.5%
26
Muse SparkMeta · Closed
89.5%
27
Claude Opus 4.6Anthropic · Closed
89.2%
28
Qwen3.8-27BAlibaba · Open weight
89.2%
29
Ornith-1.5-35B-A3BOrnith AI · Open weight
89.2%
30
DeepSeek V4 Pro (High)DeepSeek · Open weight
89.1%
31
Solar Pro 4Upstage · Closed
89.0%
32
Grok 4.20xAI · Closed
88.5%
33
DeepSeek V4 Flash 0731DeepSeek · Closed
88.1%
34
InklingThinking Machines Lab · Open weight
87.9%
35
Kimi K2.5Moonshot AI · Open weight
87.6%
36
DeepSeek V4 Flash (High)DeepSeek · Closed
87.4%
37
Hy3 PreviewTencent · Open weight
87.2%
38
MiniMax M2.7MiniMax · Open weight
87.0%
39
Nemotron 3 UltraNVIDIA · Open weight
87.0%
40
Ornith-1.5-9BOrnith AI · Open weight
86.4%
41
Solar Open 2Upstage · Open weight
86.3%
42
GLM-5.1Z.AI · Open weight
86.2%
43
GLM-5Z.AI · Open weight
86.0%
44
A.X K2SK Telecom · Open weight
85.6%
45
Ling 3.0 FlashInclusionAI · Open weight
85.0%
46
MAI-Thinking-1Microsoft · Closed
84.2%
47
Ling 3.0 Flash FP8InclusionAI · Open weight
84.0%
48
K-EXAONE 2.0LG AI Research · Open weight
82.2%
49
Mercury 2.5Inception · Closed
79.0%
50
Gemma 4 12BGoogle · Open weight
78.8%
51
Trinity-Large-ThinkingArcee AI · Open weight
76.3%
52
Nemotron 3.5 Lightning 30B A3B NVFP4NVIDIA · Open weight
75.6%
53
Mercury 2Inception · Closed
73.6%
54
DeepSeek V4 ProDeepSeek · Open weight
72.9%
55
Nemotron 3 Nano Omni 30B A3BNVIDIA · Open weight
72.2%
56
DeepSeek V4 FlashDeepSeek · Closed
71.2%
57
ZAYA1-8BZyphra · Open weight
71.0%
58
MiniCPM5-2BOpenBMB · Open weight
70.2%
59
LongCat-Flash-Lite-SparseMeituan · Open weight
69.5%
60
Trinity-Large-PreviewArcee AI · Open weight
63.3%
61
Mellum2-12B-A2.5B-ThinkingJetBrains · Open weight
57.6%
62
ZAYA1-74B-PreviewZyphra · Open weight
57.3%
63
LLaDA2.2-miniInclusionAI · Open weight
44.4%
64
Soofi S 30B-A3BSoofi Project · Open weight
43.4%
65
Mellum2-12B-A2.5B-InstructJetBrains · Open weight
40.9%
66
MiniCPM5-1BOpenBMB · Open weight
26.3%
67
LFM2.5-230MLiquidAI · Open weight
25.4%

The published GPQA-D snapshot places GPT-6 Astra first at 96.0%. The third row is 0.5 points behind. The broader top-10 range is 3.1 points, so many of the published results sit in a relatively narrow band.

67 models have been evaluated on GPQA-D. The benchmark falls in the Knowledge category. This category carries a 12% weight in BenchLM.ai's overall scoring system. GPQA-D is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About GPQA-D

Year

2026

Tasks

Graduate-level science questions

Format

Multiple choice questions

Difficulty

Graduate level

BenchLM stores GPQA-D separately from the standardized GPQA row when providers publish exact chart values that should not overwrite the core weighted benchmark.

Trinity-Large-Thinking: Scaling an Open Source Frontier Agent

BenchLM freshness & provenance

Version

GPQA-D 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.

FAQ

What does GPQA-D measure?

A display-only GPQA Diamond reference from provider comparison charts.

Which model scores highest on GPQA-D?

GPT-6 Astra by OpenAI currently leads with a score of 96.0% on GPQA-D.

How many models are evaluated on GPQA-D?

67 AI models have been evaluated on GPQA-D on BenchLM.

Compare Top Models on GPQA-D

GPT-6 Astra vs Sakana Fugu-UltraSakana Fugu-Ultra vs Sakana FuguSakana Fugu vs GPT-5.6 SolGPT-5.6 Sol vs Gemini 3.1 Pro
Last updated: September 15, 2026 · BenchLM version GPQA-D 2026

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 2,000+ readers.

One email each week. Unsubscribe anytime.