AI model benchmarks · the judgment layer
Every new AI model’s benchmark chart
says “we win.”
The maker draws it, and most of the gaps on it are noise. We read each release and tell you which differences are real — and which you should ignore.
83%
independently verified
28
AI models tracked
12
benchmarks
No composite score. Ever.
Three things a leaderboard won’t tell you — live from today’s data
Toolathlon-Verified
15.4
Real gap
Kimi K376.5
Gemini 3.1 Pro Preview61.1
Kimi K3 leads by 15.4 points on Toolathlon-Verified — outside the noise band, worth acting on.
LiveCodeBench
1.5
Tainted
Claude Opus 589.0
DeepSeek V4 Pro (0813)87.5
LiveCodeBench is saturated — rankings here are unreliable regardless of the gap.
GPQA Diamond · same model, two numbers
5.6
Why provenance matters
GLM-5.2, vendor chart91.2
GLM-5.2, independent run85.6
The maker’s own chart ran 5.6 points above the independent rerun. Every number here is tagged with who actually ran it.
Scores & pricing
All models →Why no single score? →DeepSeek V4 Pro (0813)
DeepSeek
$1.32 / $3.96
HLE
41.0independent
Terminal-Bench 2.1
78.7independent
DeepSWE
62.7self-reported
Toolathlon-Verified
74.1self-reported
Agents' Last Exam
25.7self-reported
SWE-bench Verified
96.4independent
GPQA Diamond
92.4independent
LiveCodeBench
87.5independent
ARC-AGI-2
61.3independent
LiveBench
77.4independent
HMMT
Not sourced yetAnalystAgent
Not sourced yetGPT-5.6 Sol
OpenAI
$4.00 / $20.00
HLE
49.5independent
Terminal-Bench 2.1
88.0independent
DeepSWE
73.0independent
Toolathlon-Verified
Not sourced yetAgents' Last Exam
30.6independent
SWE-bench Verified
96.2independent
GPQA Diamond
95.2independent
LiveCodeBench
82.6independent
ARC-AGI-2
92.5independent
LiveBench
81.0independent
HMMT
Not sourced yetAnalystAgent
47.5independent
Gemini 3.1 Pro Preview
Google DeepMind
$2.00 / $12.00
HLE
47.0independent
Terminal-Bench 2.1
65.8independent
DeepSWE
Not sourced yetToolathlon-Verified
61.1independent
Agents' Last Exam
16.4independent
SWE-bench Verified
78.8independent
GPQA Diamond
95.5independent
LiveCodeBench
88.5independent
ARC-AGI-2
Not sourced yetLiveBench
77.0independent
HMMT
94.7independent
AnalystAgent
41.3independent
Kimi K3
Moonshot AI
$3.00 / $15.00
HLE
46.9independent
Terminal-Bench 2.1
80.9independent
DeepSWE
69.0independent
Toolathlon-Verified
76.5independent
Agents' Last Exam
28.3independent
SWE-bench Verified
93.4independent
GPQA Diamond
92.9independent
LiveCodeBench
87.2independent
ARC-AGI-2
60.4independent
LiveBench
79.2independent
HMMT
97.0independent
AnalystAgent
38.8independent
Qwen3.8-Max
Alibaba
$2.00 / $6.00
HLE
43.1independent
Terminal-Bench 2.1
88.8independent
DeepSWE
57.0independent
Toolathlon-Verified
72.5self-reported
Agents' Last Exam
27.0independent
SWE-bench Verified
85.6independent
GPQA Diamond
93.7independent
LiveCodeBench
87.8independent
ARC-AGI-2
Not sourced yetLiveBench
78.5independent
HMMT
Not sourced yetAnalystAgent
Not sourced yetClaude Opus 5
Anthropic
$5.00 / $25.00
HLE
54.9independent
Terminal-Bench 2.1
89.1independent
DeepSWE
74.0independent
Toolathlon-Verified
Not sourced yetAgents' Last Exam
32.2independent
SWE-bench Verified
97.0independent
GPQA Diamond
93.4independent
LiveCodeBench
89.0independent
ARC-AGI-2
90.4independent
LiveBench
80.1independent
HMMT
Not sourced yetAnalystAgent
53.8independent
GLM-5.3
Zhipu AI (Z.ai)
$1.40 / $4.40
HLE
42.3independent
Terminal-Bench 2.1
83.9independent
DeepSWE
69.0independent
Toolathlon-Verified
73.0self-reported
Agents' Last Exam
Not sourced yetSWE-bench Verified
95.4independent
GPQA Diamond
91.7independent
LiveCodeBench
80.5independent
ARC-AGI-2
Not sourced yetLiveBench
76.1independent
HMMT
Not sourced yetAnalystAgent
Not sourced yetClaude Sonnet 5
Anthropic
$2.00 / $10.00
HLE
41.3independent
Terminal-Bench 2.1
74.6independent
DeepSWE
54.0independent
Toolathlon-Verified
71.6independent
Agents' Last Exam
Not sourced yetSWE-bench Verified
79.6independent
GPQA Diamond
88.9independent
LiveCodeBench
82.4independent
ARC-AGI-2
Not sourced yetLiveBench
76.0independent
HMMT
Not sourced yetAnalystAgent
46.3independent
Grok 4.6
xAI (SpaceXAI)
$2.00 / $6.00
HLE
42.9independent
Terminal-Bench 2.1
88.4independent
DeepSWE
67.0independent
Toolathlon-Verified
Not sourced yetAgents' Last Exam
Not sourced yetSWE-bench Verified
95.6independent
GPQA Diamond
94.7independent
LiveCodeBench
88.2independent
ARC-AGI-2
Not sourced yetLiveBench
78.0independent
HMMT
Not sourced yetAnalystAgent
41.3independent
MiniMax M3
MiniMax
$0.30 / $1.20
HLE
Not sourced yetTerminal-Bench 2.1
53.6independent
DeepSWE
Not sourced yetToolathlon-Verified
Not sourced yetAgents' Last Exam
Not sourced yetSWE-bench Verified
75.0independent
GPQA Diamond
92.7independent
LiveCodeBench
82.2independent
ARC-AGI-2
Not sourced yetLiveBench
67.3independent
HMMT
Not sourced yetAnalystAgent
10.0independent
GPT-5.6 Luna
OpenAI
$0.20 / $1.20
HLE
39.5independent
Terminal-Bench 2.1
80.9independent
DeepSWE
67.0independent
Toolathlon-Verified
Not sourced yetAgents' Last Exam
30.3independent
SWE-bench Verified
93.0independent
GPQA Diamond
91.7independent
LiveCodeBench
Not sourced yetARC-AGI-2
59.6independent
LiveBench
73.6independent
HMMT
Not sourced yetAnalystAgent
Not sourced yetGLM-5.3-Flash
Zhipu AI (Z.ai)
$0.15 / $0.50
HLE
39.9independent
Terminal-Bench 2.1
84.3independent
DeepSWE
63.0independent
Toolathlon-Verified
78.4self-reported
Agents' Last Exam
26.3self-reported
SWE-bench Verified
Not sourced yetGPQA Diamond
91.2independent
LiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
71.6independent
HMMT
Not sourced yetAnalystAgent
Not sourced yetQwen3.8-Flash-Next
Alibaba
—
HLE
38.0independent
Terminal-Bench 2.1
86.1independent
DeepSWE
58.7self-reported
Toolathlon-Verified
73.5self-reported
Agents' Last Exam
24.3self-reported
SWE-bench Verified
Not sourced yetGPQA Diamond
92.3independent
LiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
76.2independent
HMMT
Not sourced yetAnalystAgent
Not sourced yetClaude Fable 5.1
Anthropic
$10.00 / $50.00
HLE
59.1independent
Terminal-Bench 2.1
91.4independent
DeepSWE
Not sourced yetToolathlon-Verified
Not sourced yetAgents' Last Exam
Not sourced yetSWE-bench Verified
Not sourced yetGPQA Diamond
93.4independent
LiveCodeBench
90.5independent
ARC-AGI-2
90.0independent
LiveBench
83.4independent
HMMT
Not sourced yetAnalystAgent
57.5independent
Gemini 3.8 Flash
Google DeepMind
$0.75 / $3.75
HLE
47.8independent
Terminal-Bench 2.1
87.6independent
DeepSWE
74.0independent
Toolathlon-Verified
Not sourced yetAgents' Last Exam
Not sourced yetSWE-bench Verified
80.0independent
GPQA Diamond
94.4independent
LiveCodeBench
89.5independent
ARC-AGI-2
Not sourced yetLiveBench
Not sourced yetHMMT
Not sourced yetAnalystAgent
Not sourced yetMuse Spark 1.3
Meta (Meta Superintelligence Labs)
$1.25 / $4.25
HLE
48.7independent
Terminal-Bench 2.1
84.3independent
DeepSWE
Not sourced yetToolathlon-Verified
Not sourced yetAgents' Last Exam
32.2independent
SWE-bench Verified
Not sourced yetGPQA Diamond
93.5independent
LiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
81.6independent
HMMT
Not sourced yetAnalystAgent
Not sourced yetGPT-6 Astra
OpenAI
$10.00 / $50.00
HLE
54.7independent
Terminal-Bench 2.1
87.4independent
DeepSWE
74.0independent
Toolathlon-Verified
Not sourced yetAgents' Last Exam
34.2independent
SWE-bench Verified
Not sourced yetGPQA Diamond
96.1independent
LiveCodeBench
Not sourced yetARC-AGI-2
95.0independent
LiveBench
Not sourced yetHMMT
Not sourced yetAnalystAgent
51.3independent
DeepSeek V4.1 Flash
DeepSeek
$0.30 / $1.20
HLE
39.1self-reported
Terminal-Bench 2.1
74.5independent
DeepSWE
74.2self-reported
Toolathlon-Verified
Not sourced yetAgents' Last Exam
31.8self-reported
SWE-bench Verified
Not sourced yetGPQA Diamond
90.9self-reported
LiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
Not sourced yetHMMT
Not sourced yetAnalystAgent
Not sourced yetStep 5 Preview
StepFun
$1.00 / $2.70
HLE
46.5independent
Terminal-Bench 2.1
85.0self-reported
DeepSWE
67.7self-reported
Toolathlon-Verified
74.1self-reported
Agents' Last Exam
Not sourced yetSWE-bench Verified
Not sourced yetGPQA Diamond
93.5self-reported
LiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
Not sourced yetHMMT
Not sourced yetAnalystAgent
Not sourced yetMiMo-V2.6-Pro
Xiaomi
$0.43 / $0.87
HLE
49.4independent
Terminal-Bench 2.1
67.8independent
DeepSWE
71.9self-reported
Toolathlon-Verified
76.9self-reported
Agents' Last Exam
31.6self-reported
SWE-bench Verified
Not sourced yetGPQA Diamond
Not sourced yetLiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
Not sourced yetHMMT
Not sourced yetAnalystAgent
Not sourced yetMiMo-V2.6-Flash
Xiaomi
$0.14 / $0.28
HLE
Not sourced yetTerminal-Bench 2.1
76.4independent
DeepSWE
67.9self-reported
Toolathlon-Verified
73.6self-reported
Agents' Last Exam
27.6self-reported
SWE-bench Verified
Not sourced yetGPQA Diamond
Not sourced yetLiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
Not sourced yetHMMT
Not sourced yetAnalystAgent
Not sourced yetGrok 4.7
xAI (SpaceXAI)
$2.00 / $6.00
HLE
43.1independent
Terminal-Bench 2.1
73.4independent
DeepSWE
71.0self-reported
Toolathlon-Verified
Not sourced yetAgents' Last Exam
Not sourced yetSWE-bench Verified
Not sourced yetGPQA Diamond
Not sourced yetLiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
77.4independent
HMMT
Not sourced yetAnalystAgent
Not sourced yetClaude Opus 5.5
Anthropic
$4.00 / $20.00
HLE
61.4independent
Terminal-Bench 2.1
87.6independent
DeepSWE
Not sourced yetToolathlon-Verified
Not sourced yetAgents' Last Exam
38.2independent
SWE-bench Verified
Not sourced yetGPQA Diamond
Not sourced yetLiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
83.2independent
HMMT
Not sourced yetAnalystAgent
Not sourced yetGPT-6 Sol
OpenAI
$2.00 / $10.00
HLE
47.9independent
Terminal-Bench 2.1
83.2independent
DeepSWE
69.0independent
Toolathlon-Verified
Not sourced yetAgents' Last Exam
32.2independent
SWE-bench Verified
Not sourced yetGPQA Diamond
Not sourced yetLiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
79.3independent
HMMT
Not sourced yetAnalystAgent
Not sourced yetGPT-6 Luna
OpenAI
$0.10 / $0.50
HLE
38.5independent
Terminal-Bench 2.1
73.0independent
DeepSWE
66.6self-reported
Toolathlon-Verified
Not sourced yetAgents' Last Exam
25.0independent
SWE-bench Verified
Not sourced yetGPQA Diamond
Not sourced yetLiveCodeBench
Not sourced yetARC-AGI-2
59.3independent
LiveBench
72.0independent
HMMT
Not sourced yetAnalystAgent
Not sourced yetGPT-6.1 Sol
OpenAI
$2.00 / $10.00
HLE
52.9independent
Terminal-Bench 2.1
Not sourced yetDeepSWE
69.6independent
Toolathlon-Verified
Not sourced yetAgents' Last Exam
Not sourced yetSWE-bench Verified
Not sourced yetGPQA Diamond
Not sourced yetLiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
81.6independent
HMMT
Not sourced yetAnalystAgent
Not sourced yetTencent Hy4 preview
Tencent
$0.83 / $2.50
HLE
43.4self-reported
Terminal-Bench 2.1
55.1independent
DeepSWE
64.3self-reported
Toolathlon-Verified
74.1self-reported
Agents' Last Exam
Not sourced yetSWE-bench Verified
Not sourced yetGPQA Diamond
92.3self-reported
LiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
Not sourced yetHMMT
Not sourced yetAnalystAgent
Not sourced yetClaude Sonnet 5.5
Anthropic
$2.00 / $10.00
HLE
55.0independent
Terminal-Bench 2.1
83.2independent
DeepSWE
71.0self-reported
Toolathlon-Verified
77.8self-reported
Agents' Last Exam
Not sourced yetSWE-bench Verified
Not sourced yetGPQA Diamond
Not sourced yetLiveCodeBench
Not sourced yetARC-AGI-2
Not sourced yetLiveBench
77.8independent
HMMT
Not sourced yetAnalystAgent
Not sourced yet| Pricein / out per 1M | Reasoning±2 is noise | Terminal ops±10.6 is noise | Long-horizon coding±9.5 is noise | Multi-tool chores±9.7 is noise | Professional work±3.2 is noise | Bug fixingsaturated · not ranked | Expert science Q&Asaturated · not ranked | Contest codingsaturated · not ranked | Compositional visual reasoning±9.2 is noise | Composite score across 7 domains±2.7 is noise | Competition mathematicscontaminated · not ranked | Spreadsheet & document analysis±11.2 is noise | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V4 Pro (0813) DeepSeek | $1.32 / $3.96 | 41.0independent | 78.7independent | 62.7self-reported | 74.1self-reported | 25.7self-reported | 96.4independent | 92.4independent | 87.5independent | 61.3independent | 77.4independent | Not sourced yet | Not sourced yet |
| GPT-5.6 Sol OpenAI | $4.00 / $20.00 | 49.5independent | 88.0independent | 73.0independent | Not sourced yet | 30.6independent | 96.2independent | 95.2independent | 82.6independent | 92.5independent | 81.0independent | Not sourced yet | 47.5independent |
| Gemini 3.1 Pro Preview Google DeepMind | $2.00 / $12.00 | 47.0independent | 65.8independent | Not sourced yet | 61.1independent | 16.4independent | 78.8independent | 95.5independent | 88.5independent | Not sourced yet | 77.0independent | 94.7independent | 41.3independent |
| Kimi K3 Moonshot AI | $3.00 / $15.00 | 46.9independent | 80.9independent | 69.0independent | 76.5independent | 28.3independent | 93.4independent | 92.9independent | 87.2independent | 60.4independent | 79.2independent | 97.0independent | 38.8independent |
| Qwen3.8-Max Alibaba | $2.00 / $6.00 | 43.1independent | 88.8independent | 57.0independent | 72.5self-reported | 27.0independent | 85.6independent | 93.7independent | 87.8independent | Not sourced yet | 78.5independent | Not sourced yet | Not sourced yet |
| Claude Opus 5 Anthropic | $5.00 / $25.00 | 54.9independent | 89.1independent | 74.0independent | Not sourced yet | 32.2independent | 97.0independent | 93.4independent | 89.0independent | 90.4independent | 80.1independent | Not sourced yet | 53.8independent |
| GLM-5.3 Zhipu AI (Z.ai) | $1.40 / $4.40 | 42.3independent | 83.9independent | 69.0independent | 73.0self-reported | Not sourced yet | 95.4independent | 91.7independent | 80.5independent | Not sourced yet | 76.1independent | Not sourced yet | Not sourced yet |
| Claude Sonnet 5 Anthropic | $2.00 / $10.00 | 41.3independent | 74.6independent | 54.0independent | 71.6independent | Not sourced yet | 79.6independent | 88.9independent | 82.4independent | Not sourced yet | 76.0independent | Not sourced yet | 46.3independent |
| Grok 4.6 xAI (SpaceXAI) | $2.00 / $6.00 | 42.9independent | 88.4independent | 67.0independent | Not sourced yet | Not sourced yet | 95.6independent | 94.7independent | 88.2independent | Not sourced yet | 78.0independent | Not sourced yet | 41.3independent |
| MiniMax M3 MiniMax | $0.30 / $1.20 | Not sourced yet | 53.6independent | Not sourced yet | Not sourced yet | Not sourced yet | 75.0independent | 92.7independent | 82.2independent | Not sourced yet | 67.3independent | Not sourced yet | 10.0independent |
| GPT-5.6 Luna OpenAI | $0.20 / $1.20 | 39.5independent | 80.9independent | 67.0independent | Not sourced yet | 30.3independent | 93.0independent | 91.7independent | Not sourced yet | 59.6independent | 73.6independent | Not sourced yet | Not sourced yet |
| GLM-5.3-Flash Zhipu AI (Z.ai) | $0.15 / $0.50 | 39.9independent | 84.3independent | 63.0independent | 78.4self-reported | 26.3self-reported | Not sourced yet | 91.2independent | Not sourced yet | Not sourced yet | 71.6independent | Not sourced yet | Not sourced yet |
| Qwen3.8-Flash-Next Alibaba | — | 38.0independent | 86.1independent | 58.7self-reported | 73.5self-reported | 24.3self-reported | Not sourced yet | 92.3independent | Not sourced yet | Not sourced yet | 76.2independent | Not sourced yet | Not sourced yet |
| Claude Fable 5.1 Anthropic | $10.00 / $50.00 | 59.1independent | 91.4independent | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | 93.4independent | 90.5independent | 90.0independent | 83.4independent | Not sourced yet | 57.5independent |
| Gemini 3.8 Flash Google DeepMind | $0.75 / $3.75 | 47.8independent | 87.6independent | 74.0independent | Not sourced yet | Not sourced yet | 80.0independent | 94.4independent | 89.5independent | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet |
| Muse Spark 1.3 Meta (Meta Superintelligence Labs) | $1.25 / $4.25 | 48.7independent | 84.3independent | Not sourced yet | Not sourced yet | 32.2independent | Not sourced yet | 93.5independent | Not sourced yet | Not sourced yet | 81.6independent | Not sourced yet | Not sourced yet |
| GPT-6 Astra OpenAI | $10.00 / $50.00 | 54.7independent | 87.4independent | 74.0independent | Not sourced yet | 34.2independent | Not sourced yet | 96.1independent | Not sourced yet | 95.0independent | Not sourced yet | Not sourced yet | 51.3independent |
| DeepSeek V4.1 Flash DeepSeek | $0.30 / $1.20 | 39.1self-reported | 74.5independent | 74.2self-reported | Not sourced yet | 31.8self-reported | Not sourced yet | 90.9self-reported | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet |
| Step 5 Preview StepFun | $1.00 / $2.70 | 46.5independent | 85.0self-reported | 67.7self-reported | 74.1self-reported | Not sourced yet | Not sourced yet | 93.5self-reported | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet |
| MiMo-V2.6-Pro Xiaomi | $0.43 / $0.87 | 49.4independent | 67.8independent | 71.9self-reported | 76.9self-reported | 31.6self-reported | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet |
| MiMo-V2.6-Flash Xiaomi | $0.14 / $0.28 | Not sourced yet | 76.4independent | 67.9self-reported | 73.6self-reported | 27.6self-reported | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet |
| Grok 4.7 xAI (SpaceXAI) | $2.00 / $6.00 | 43.1independent | 73.4independent | 71.0self-reported | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | 77.4independent | Not sourced yet | Not sourced yet |
| Claude Opus 5.5 Anthropic | $4.00 / $20.00 | 61.4independent | 87.6independent | Not sourced yet | Not sourced yet | 38.2independent | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | 83.2independent | Not sourced yet | Not sourced yet |
| GPT-6 Sol OpenAI | $2.00 / $10.00 | 47.9independent | 83.2independent | 69.0independent | Not sourced yet | 32.2independent | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | 79.3independent | Not sourced yet | Not sourced yet |
| GPT-6 Luna OpenAI | $0.10 / $0.50 | 38.5independent | 73.0independent | 66.6self-reported | Not sourced yet | 25.0independent | Not sourced yet | Not sourced yet | Not sourced yet | 59.3independent | 72.0independent | Not sourced yet | Not sourced yet |
| GPT-6.1 Sol OpenAI | $2.00 / $10.00 | 52.9independent | Not sourced yet | 69.6independent | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | 81.6independent | Not sourced yet | Not sourced yet |
| Tencent Hy4 preview Tencent | $0.83 / $2.50 | 43.4self-reported | 55.1independent | 64.3self-reported | 74.1self-reported | Not sourced yet | Not sourced yet | 92.3self-reported | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet |
| Claude Sonnet 5.5 Anthropic | $2.00 / $10.00 | 55.0independent | 83.2independent | 71.0self-reported | 77.8self-reported | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | Not sourced yet | 77.8independent | Not sourced yet | Not sourced yet |
INDEPENDENT = a third party ran the benchmark itself (tbench.ai, vals.ai, Artificial Analysis, Snorkel…).
SELF-REPORTED = the vendor’s own number, not yet replicated.
Every cell links to its source — methodology.
There is no total score. We won’t collapse 12 different benchmarks into one number that pretends a 0.5-point gap means something — see a head-to-head for which gaps here are real.