Model profile
DeepSeek
New in 2026

DeepSeek: DeepSeek V4 Pro

DeepSeek: DeepSeek V4 Pro is a budget-priced text-first model from DeepSeek with heavy runtime profile, large context posture, and the clearest fit around long-context research / agent workflows.

Best for: Long-context research / Agent workflowsHeavy latencyLarge contextBudget pricing
Intelligence
51.5

Benchmark blend

Coding
47.5

Dev workflow signal

Context
1049K Tokens

Large

Input Price
$1.74

Budget tier

Decision snapshot
63

DeepSeek: DeepSeek V4 Pro currently reads as a budget text-first option with large context and a heavy runtime profile.

Overall profile
Selective fit
Best for
Long-context research / Agent workflows
Latency tier
Heavy
Price tier
Budget
Source coverage
OpenRouterArtificial Analysis

Decision Strip

Decision rail before the raw tables

Core buy-side signals stay in one pass. The rest of the page expands only after intelligence, speed, context, and price are clear.

Intelligence
51.5
52

General reasoning and benchmark headroom.

Situational
Speed
38 tok/s
47

TTFT 1.59s

Situational
Context
1049K Tokens
100

How much prompt and task state can stay in view.

Above average
Price
$1.74
86

$3.48 output / 1M

Efficient

Editorial Profile

DeepSeek: DeepSeek V4 Pro in one narrative

Positioning, tradeoffs, and fit are consolidated into one read instead of repeating the same story across separate cards.

Selective fitCoding score 48Math score N/A

DeepSeek V4 Pro is a large-scale Mixture-of-Experts model from DeepSeek with 1.6T total parameters and 49B activated parameters, supporting a 1M-token context window. It is designed for advanced reasoning, coding,...

Identity

DeepSeek text-first profile

Positioning

Long-context research / Agent workflows with large context and heavy runtime.

Cost posture

Efficient spend profile. More comfortable for sustained prompt volume if the capability fit is right.

Strengths
  • Large context headroom supports repo-wide prompts and long research sessions.

Tradeoffs
  • Budget-friendly input pricing is a strength, but raw capability may vary by workload.

  • Latency profile is better for deliberate runs than rapid back-and-forth chat.

  • Current metadata points to a text-first profile rather than a broad multimodal one.

Best fit
  • Long-context summarization, repo analysis, and policy or document review.

Explore Next

Similar profiles worth opening next

DeepSeek

DeepSeek V4 Pro (Reasoning, High Effort)

Intelligence
49.8
Context
N/A
Input Price
$1.74
DeepSeek

DeepSeek: DeepSeek V4 Flash

Intelligence
46.5
Context
1049K Tokens
Input Price
$0.14
DeepSeek

DeepSeek V4 Flash (Reasoning, High Effort)

Intelligence
44.9
Context
N/A
Input Price
$0.14

Benchmarks

Grouped by job-to-be-done

Only benchmark categories with actual signal are shown. Secondary values stay as simple definitions instead of nested micro-cards.

General intelligence

Broad reasoning, knowledge depth, and flagship benchmark posture.

Intelligence Index
51.5
GPQA
88.8%
HLE
35.9%
Coding

Software implementation, debugging quality, and coding benchmark signal.

Coding Index
47.5
SciCode
50.0%
Agent / tool use

Long-horizon execution quality and interactive benchmark evidence.

IFBench
76.5%
TAU2
96.2%
TerminalBench Hard
46.2%
LCR
66.3%

Specs & Pricing

Technical snapshot and cost posture

Specs stay neutral, pricing gets emphasis through values rather than extra containers. Raw provider internals remain in metadata at the end.

Technical snapshot
Context Window
1049K Tokens
Vision
Text-first
Modalities
text
Tokenizer
DeepSeek
Max Completion
384000
Moderation
No
Supported Parameters
frequency_penaltyinclude_reasoninglogprobsmax_tokenspresence_penaltyreasoningresponse_formatseedstoptemperaturetool_choicetoolstop_ktop_logprobstop_p
Input Modalities
text
Output Modalities
text
Price architecture
Input
per 1M input tokens
$1.74
Output
per 1M output tokens
$3.48
Blended
AA 3:1 mix
$2.17

This model is relatively efficient on price. It is the easier fit when sustained prompt volume matters.

OR Cache Read
$0.00

Metadata

Raw source tables at the end

Verification details remain available, but the page no longer forces them ahead of the editorial read.