Model comparison
DeepSeek V4-Flash vs DeepSeek V4-Pro
Two DeepSeek tiers compared on the figures that decide which one a workload actually needs.
Catalog record checked August 14, 2026; individual provider fields may change.
DeepSeek
DeepSeek V4-Flash
Budget · Open weights
DeepSeek
DeepSeek V4-Pro
Balanced · Open weights
| Specification | DeepSeek V4-Flash | DeepSeek V4-Pro |
|---|---|---|
| Provider | DeepSeek | DeepSeek |
| Tier | Budget | Balanced |
| Context window | 1M | 1M |
| Max output | 384K | 384K |
| Input / 1M tokens | Winner: $0.14 | $0.435 |
| Output / 1M tokens | Winner: $0.28 | $0.87 |
| Weights | Open | Open |
| Parameters | 284B total / 13B active (MoE) | 1.6T total / 49B active (MoE) |
| Reasoning levels | low, high, max | low, high, max |
| Modalities | text | text |
| Released | July 31, 2026 | April 24, 2026 |
| Artificial Analysis Intelligence Index [max] (2026-08-08) | 52 | Winner: 53 |
Prices are USD per million tokens at standard rates, excluding batch and caching discounts. Bold indicates the better figure where one is objectively better. Values we could not confirm from the provider are shown as “Not verified” rather than estimated.
Pricing tiers: DeepSeek V4-Flash: Current list is $0.14/$0.28 per million tokens (cache miss / output) until 2026-08-16 16:00 UTC. From then, official peak/off-peak rates: off-peak $0.22/$0.66, peak $0.44/$1.32 (cache miss / output). Peak hours are 01:00–04:00 and 06:00–10:00 UTC. · DeepSeek V4-Pro: API id `deepseek-v4-pro` is unchanged across the 2026-04-24 preview and the 2026-08-13 GA checkpoint. Current list is $0.435/$0.87 per million tokens (cache miss / output) until 2026-08-16 16:00 UTC. From then, official peak/off-peak rates: off-peak $0.66/$1.98, peak $1.32/$3.96 (cache miss / output). Peak hours are 01:00–04:00 and 06:00–10:00 UTC.
DeepSeek V4-Flash
Open-weight price-performance pick: MIT weights, 1M context, and API rates far below closed budget tiers.
Best for
- Cost-sensitive hosted agents
- High-volume coding assist
- Open-weight deployments with cluster VRAM
Watch out
API list prices move to peak/off-peak from 2026-08-16 16:00 UTC (see pricing note). Self-hosting the ~284B MoE still needs roughly 90–170GB+ class VRAM depending on quant — not a laptop budget build. DeepSWE snapshot also uses the shared mini-swe-agent harness — verify cost, effort, and serving conditions before treating the result as a forecast.
DeepSeek V4-Pro
DeepSeek's larger V4 sibling for harder reasoning and coding, still priced well below closed mid tiers.
Best for
- Harder reasoning on a budget
- Open-weight deployments
- Agentic coding when Flash is not enough
Watch out
The April preview and 2026-08-13 GA share one API id — pin the date of any score you cite. Artificial Analysis currently evaluates the 0813 (max) checkpoint at Intelligence Index 53, Terminal-Bench v2.1 78.7%, and GPQA Diamond 92.8%. Peak/off-peak API rates take effect 2026-08-16 16:00 UTC.
Benchmark
DeepSWE 1.1 in context
Both models shown against the wider field, with cost per completed task alongside the score.
Local leader
Claude Opus 5 [max]
74%
Rows shown
24
Highest published reasoning effort per model (not best Pass@1)
Snapshot date
2026-08-13
Mirrored from deepswe.datacurve.ai
Better is toward the top-right (higher pass rate, lower cost). X-axis is reversed to match DeepSWE’s public chart. v1.1 uses average cost / tokens / steps; v1 uses published medians.
| # | Model | Pass@1 | Cost / task | Tokens / task | Steps / task |
|---|---|---|---|---|---|
| 1 | Claude Opus 5 [max] | 74% | $11.84 | 118k | 99 |
| 2 | GPT-5.6 Sol [max] | 73% | $8.39 | 60k | 61 |
| 3 | Claude Fable 5 [max] | 70% | $21.63 | 119k | 88 |
| 4 | GPT-5.6 Terra [max] | 70% | $4.95 | 72k | 76 |
| 5 | Kimi K3 [max] | 69% | $4.65 | 82k | 98 |
| 6 | GPT-5.6 Luna [max] | 67% | $3.03 | 73k | 102 |
| 7 | GPT-5.5 [xhigh] | 67% | $7.23 | 46k | 82 |
| 8 | Grok 4.6 [xhigh] | 67% | $5.50 | 71k | 87 |
| 9 | Gemini 3.7 Flash [high] | 65% | $2.18 | 107k | 125 |
| 10 | DeepSeek V4-Pro [max] | 63% | $0.24 | 106k | 155 |
| 11 | Claude Opus 4.8 [max] | 59% | $13.22 | 135k | 120 |
| 12 | Qwen3.8-Max [xhigh] | 58% | $3.73 | 95k | 111 |
| 13 | Muse Spark 1.2 [xhigh] | 55% | $3.70 | 99k | 101 |
| 14 | Claude Sonnet 5 [max] | 54% | $26.40 | 214k | 268 |
| 15 | Grok 4.5 [high] | 54% | $2.42 | 36k | 61 |
| 16 | DeepSeek V4-Flash [max] | 53% | $0.10 | 108k | 153 |
| 17 | Muse Spark 1.1 [xhigh] | 53% | $2.36 | 74k | 96 |
| 18 | GPT-5.4 [xhigh] | 52% | $5.65 | 71k | 70 |
| 19 | Gemini 3.6 Flash [high] | 47% | $4.42 | 96k | 117 |
| 20 | GLM 5.2 [max] | 44% | $3.92 | 78k | 129 |
| 21 | Gemini 3.5 Flash [high] | 36% | $3.45 | 76k | 105 |
| 22 | Kimi K2.7 Code | 31% | $2.82 | 59k | 149 |
| 23 | Claude Sonnet 4.6 [high] | 30% | $5.52 | 76k | 134 |
| 24 | Gemini 3.1 Pro [high] | 12% | $2.14 | 28k | 76 |
DeepSWE “Best” picks the highest published reasoning effort per model (not the highest pass rate). Small gaps may not be statistically meaningful — confirm on deepswe.datacurve.ai.
Common questions
DeepSeek V4-Flash vs DeepSeek V4-Pro
Answered from the verified figures on this page rather than general guidance.
Is DeepSeek V4-Flash or DeepSeek V4-Pro cheaper for input?
DeepSeek V4-Flash is cheaper at $0.14 per million input tokens, against $0.435 for DeepSeek V4-Pro — roughly 3.1× the price. Output tokens usually dominate a real bill, so weigh the output rate more heavily than the input rate. These are base rates; DeepSeek V4-Flash has tiered pricing: Current list is $0.14/$0.28 per million tokens (cache miss / output) until 2026-08-16 16:00 UTC. From then, official peak/off-peak rates: off-peak $0.22/$0.66, peak $0.44/$1.32 (cache miss / output). Peak hours are 01:00–04:00 and 06:00–10:00 UTC. DeepSeek V4-Pro has tiered pricing: API id `deepseek-v4-pro` is unchanged across the 2026-04-24 preview and the 2026-08-13 GA checkpoint. Current list is $0.435/$0.87 per million tokens (cache miss / output) until 2026-08-16 16:00 UTC. From then, official peak/off-peak rates: off-peak $0.66/$1.98, peak $1.32/$3.96 (cache miss / output). Peak hours are 01:00–04:00 and 06:00–10:00 UTC.
Is DeepSeek V4-Flash or DeepSeek V4-Pro cheaper for output?
DeepSeek V4-Flash is cheaper at $0.28 per million output tokens, against $0.87 for DeepSeek V4-Pro — roughly 3.1× the price. Output tokens usually dominate a real bill, so weigh the output rate more heavily than the input rate. These are base rates; DeepSeek V4-Flash has tiered pricing: Current list is $0.14/$0.28 per million tokens (cache miss / output) until 2026-08-16 16:00 UTC. From then, official peak/off-peak rates: off-peak $0.22/$0.66, peak $0.44/$1.32 (cache miss / output). Peak hours are 01:00–04:00 and 06:00–10:00 UTC. DeepSeek V4-Pro has tiered pricing: API id `deepseek-v4-pro` is unchanged across the 2026-04-24 preview and the 2026-08-13 GA checkpoint. Current list is $0.435/$0.87 per million tokens (cache miss / output) until 2026-08-16 16:00 UTC. From then, official peak/off-peak rates: off-peak $0.66/$1.98, peak $1.32/$3.96 (cache miss / output). Peak hours are 01:00–04:00 and 06:00–10:00 UTC.
Which has the larger context window, DeepSeek V4-Flash or DeepSeek V4-Pro?
Both accept about 1M tokens of context, so document length will not decide between them.
Do DeepSeek V4-Flash and DeepSeek V4-Pro support the same reasoning levels?
Yes — both accept the same effort settings: "low", "high", "max". Higher effort costs more and takes longer, so start low and raise it only where output quality actually improves.
Should I use DeepSeek V4-Flash or DeepSeek V4-Pro?
DeepSeek V4-Flash is the budget tier and DeepSeek V4-Pro the balanced tier. The useful question is whether your hardest task actually fails on the cheaper one — most production volume such as classification, extraction and summarisation does not.
Next step
Choosing between them
Tier and workload decide this more reliably than a leaderboard position does.
If both sit in the same tier, the decision usually comes down to context window and output price rather than headline capability — output tokens dominate real bills.
If one is a step up within the same provider, the useful question is whether your hardest task actually fails on the cheaper tier. Most production volume — classification, extraction, summarization — does not.
Choosing a harness rather than a model — Cursor, Copilot, Claude Code, Muse Code, or Lovable? Compare agentic harnesses · Latest releases.