none reviewed yet · 4 benchmark observations · latest capture Sep 8, 2026
Inspect the evidenceDeepSeek · MODEL PROFILE
DeepSeek V4 Flash
High-concurrency, low-latency coding assistance and batch page generation on a tight budget
Saved editorial fit for your current task lens; not a benchmark result
How scoring worksVerified rates (Sep 8, 2026) · 100 runs / month at your token settings
Adjust workloadAt a glance
9 facts- Provider
- DeepSeek
- Type
- API model
- Access
- DeepSeek API (OpenAI- and Anthropic-compatible), Alibaba Cloud Model Studio, OpenRouter, self-hosting; concurrency limit 2,500 vs 500 for V4 Pro
- Context
- 1,048,576 tokens
- Maximum output
- 384K tokens
- License / weights
- Open weights
- Input modalities
- text
- Output modalities
- text
- Pricing notes
- Peak $0.44 in / $1.32 out per 1M; off-peak exactly half at $0.22/$0.66; cache-hit input $0.014 peak / $0.007 off-peak. Peak hours 01:00-04:00 and 06:00-10:00 UTC Mon-Fri.
The practical trade-offs
Documented strengths
- Fastest time-to-first-token measured here (0.95 s) with 131.8 tok/s throughput
- Terminal-Bench 2.1 82.7% and SWE-bench Verified 79.0% at $0.44/$1.32
- MIT-licensed 304B weights - practical to self-host on a single high-end node
- 2,500-request concurrency limit, 5x the V4 Pro tier
Limitations to check
- Design Arena Website 1223-1255 puts visual taste in the bottom third
- SWE-bench Pro 52.6% is mid-pack
- Text-only unless you switch to the vision-exp sibling
- Two different Design Arena rows for near-identical names invite confusion
Follow the source
9 source linksEvidence & conditions
9Source links
NoneReviewed records
4Benchmark runs
Sep 8, 2026Latest capture
- Evidence source 1
api-docs.deepseek.comDate not recorded - Evidence source 2
huggingface.coDate not recorded - Evidence source 3
arena.aiDate not recorded - Evidence source 4
benchlm.aiDate not recorded - Evidence source 5
benchlm.aiDate not recorded - Evidence source 6
benchlm.aiDate not recorded - Evidence source 7
artificialanalysis.aiDate not recorded - Evidence source 8
digitalapplied.comDate not recorded - Evidence source 9
docs.qwencloud.comDate not recorded
Benchmark observations
Terminal-Bench 2.1 Inspect conditions
- name
- Terminal-Bench 2.1
- value
- 82.7
- unit
- %
- source Url
- https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731
- checked At
- 2026-09-08T19:28:18Z
- attribution
- provider-reported
- configuration
- Exact DeepSeek-V4-Flash-0731; minimal DeepSeek Harness; reasoning_effort=max; temperature=1; top_p=0.95.
DeepSWE Inspect conditions
- name
- DeepSWE
- value
- 54.4
- unit
- %
- source Url
- https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731
- checked At
- 2026-09-08T19:28:18Z
- attribution
- provider-reported
- configuration
- Exact DeepSeek-V4-Flash-0731; minimal DeepSeek Harness; reasoning_effort=max; temperature=1; top_p=0.95.
NL2Repo Inspect conditions
- name
- NL2Repo
- value
- 54.2
- unit
- %
- source Url
- https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731
- checked At
- 2026-09-08T19:28:18Z
- attribution
- provider-reported
- configuration
- Exact DeepSeek-V4-Flash-0731; minimal DeepSeek Harness; reasoning_effort=max; temperature=1; top_p=0.95.
Arena WebDev Frontend Inspect conditions
- name
- Arena WebDev Frontend
- value
- 1582
- unit
- arena_score
- source Url
- https://arena.ai/leaderboard/code/webdev/frontend
- checked At
- 2026-09-08T19:28:18Z
- attribution
- benchmark-publisher
- configuration
- Exact row deepseek-v4-flash-high; source results dated 2026-09-05; human preference; confidence interval +/-11; 3377 votes. This is a checked source snapshot, not a newly run September 8 evaluation.
Source reports, measured results, and editorial assessments are separate evidence types. Source dates do not imply a new benchmark run. Capture dates are shown when recorded; otherwise we say the capture date was not recorded. Recorded dates are compared with this catalogue snapshot (Sep 11, 2026), not today’s date.
Method AIS-V2
How this page is assessed
Quality claims follow Method AIS-V2.
Money never buys rank — commercial relationships cannot change rank, eligibility, evidence quality, or publication timing.
Submit or claim a listing is free and does not skip the editorial queue.
Private · this device only