CADLE CAP
94
Cadle #19/385 · 82% confidence
DeepSeek · Open weightsDeepSeek: DeepSeek V4 Pro 0813
Active
Model portfolio
DeepSeek: DeepSeek V4 Pro 0813
Tap a star to rate
DeepSeek · Released 2026-08-12 · v2026-08-12 · Open · Context 1.0M tokens
DeepSeek: DeepSeek V4 Pro 0813 is an open-source model from DeepSeek (released 2026-08-12). Context window ~1.0M tokens with capabilities: Text, Reasoning. Positioned for higher-effort reasoning workloads. List pricing ≈ $0.66 / $1.98 per 1M input/output tokens. Top tracked signals: MATH · epoch 98.6, MATH 96.9, CODING · benchlm 96.4. Cadle merges provider listings, unified leaderboard medians, and public pricing where available.
Language Cap
94
Cadle Language #19/385 · 82%
Benchmark average
83.1
#46 on the benchmark matrix
Pricing
$0.66
$1.98 / 1M out · 32.0 pts/$
CROSS-BOARD RANKS
Cadle Language LLM #19/385Matrix #46/443BenchLM #43/224AA #88/698vs BenchLM: -25
CADLE CAP vs MODEL MATRIX
Cadle Cap ranks models on every published score we have for them — harder benchmarks weigh more. Missing tests are skipped, not zeroed. It answers who is stronger than whom in today's pool.
Model Matrix is the datasheet: Overall, Intelligence, Coding, Agentic, MMLU, GPQA, Math, LiveCodeBench, HumanEval, Reason, ARC-AGI. A dash means unmeasured. AVG is an importance-weighted mean of that model's filled cells only.
MODEL MATRIX · TRUTH CENTER CELLS
Overall
63.1
AA IQ
36.0
Coding
82.6
Agentic
87.4
MMLU
94.1
GPQA
92.8
Math
96.9
LiveCodeBench
72.3
HumanEval
—
Reason
90.0
ARC-AGI
61.3
Coverage AVG
83.1
Cadle Cap ranks models with within-benchmark percentiles so ARC-AGI ~40 and MMLU ~90 never compete raw. Coverage AVG is an importance-weighted mean of filled expert cells only — blanks are skipped, not zeroed.
ANALYSIS · TRUTH CENTER
Hard benches (GPQA · ARC-AGI · LCB · Agentic) sit left — a mid-40 on ARC-AGI can still be frontier. Broader benches on the right are easier and often higher.
CADLE INSIGHTS
Only measured signals
INTELLIGENCE
45.7
Media Cap quality axis
CODING
82.6
Media Cap quality axis
AGENTIC
87.4
Media Cap quality axis
MATH
96.9
Media Cap quality axis
MATRIX OVERALL
63.1
AVG 83.1 across matrix cells
BENCHMARK AVG
79.9
26 tracked scores
STRONGEST SIGNAL
MATH · epoch
Score 98.6
SPEED · COST · QUALITY
Measured matrix / AA / gateway rows only.
- Overall (Matrix)
- 63.1
- Cap Intelligence
- 84.5
- AA / Intel index
- 36.0
- Cap Coding
- 97.6
- Matrix Coding
- 82.6
- Input $/1M
- $0.66
- Output $/1M
- $1.98
- Value pts/$
- 31.99
DESCRIPTION + DETAILS
DeepSeek: DeepSeek V4 Pro 0813 is an open-source model from DeepSeek (released 2026-08-12). Context window ~1.0M tokens with capabilities: Text, Reasoning. Positioned for higher-effort reasoning workloads. List pricing ≈ $0.66 / $1.98 per 1M input/output tokens. Top tracked signals: MATH · epoch 98.6, MATH 96.9, CODING · benchlm 96.4. Cadle merges provider listings, unified leaderboard medians, and public pricing where available.
PROVIDER
DeepSeek
RELEASE
2026-08-12
CONTEXT
1.0M tokens
STATUS
ACTIVE
COMMUNITY REVIEWS
OPERATOR NOTES
Quality, latency, cost, and failure modes — not marketing copy.
No reviews yet
WRITE A REVIEW
Tap a star to rate
Stored on this device · synced when signed in
No operator notes yet
First review sets the tone for how teams evaluate DeepSeek: DeepSeek V4 Pro 0813.
MODEL COMPARISON
| Metric | |
|---|---|
| Provider | DeepSeek |
| Release | 2026-08-12 |
| Input / 1M | $0.66 |
| Output / 1M | $1.98 |
| Context | 1.0M tokens |
| Cadle Cap | 94 |
| Cap board | Cadle Language #19/385 |
| MATH · epoch | 98.6 |
| MATH | 96.9 |
| CODING · benchlm | 96.4 |
| AGENTIC · benchlm | 96.2 |
| MATH · benchlm | 95.2 |
| MMLU · benchlm | 94.1 |
| MMLU | 94.1 |
| LIVECODEBENCH · benchlm | 93.5 |
| GPQA · benchlm | 92.8 |
| GPQA | 92.8 |
| GPQA · artificial-analysis | 92.8 |
| GPQA · epoch | 92.6 |
DATA PROVENANCE
Scores may use different scales — check the Benchmarks tab for source links.
- OpenRouter / provider APIs (pricing & context)
- Hugging Face Hub (downloads & cards)
- Papers With Code (task benchmarks)
- BenchLM.ai (category rollups + atomic benches)
- Epoch AI Benchmarking Hub (ECI + ZIP CSVs)
- LiveCodeBench (generation / execution / repair / test-output)
- Artificial Analysis (intelligence / indexes)
- LLM Stats (leaderboard + rankings)
- Cadle unified median (equal-weight multi-source)
- Epoch AI Benchmarking Hub — capability CSV exports (when synced)
SOURCE TRACKING
Evidence linked to this model — listing, hub, benchmarks, or live gateway prices.
API
DeepSeek listing
API
OpenRouter
COMMUNITY
Papers with Code
COMMUNITY
BenchLM
COMMUNITY
epoch
COMMUNITY
unified-median
COMMUNITY
artificial-analysis
MEASURED CELLS
Same Truth Center numbers as the datasheet above. Accuracy, reliability, safety, and cost scores are not invented.
Overall63.1
AA IQ36
Coding82.6
Agentic87.4
MMLU94.1
GPQA92.8
Math96.9
LiveCodeBench72.3
ARC-AGI61.3