Benchmarks
+Where deepseek-v4-pro and deepseek-v4-flash
+sit against the field, what the 0813 checkpoint changed, and the economics
+argument the Chinese community calls the 斩杀线, the
+kill line. The numbers below are DeepSeek's own launch-day figures unless
+marked otherwise; read the caveats first.
-
+
- Vendor numbers, one harness. The table is DeepSeek's own +agent-benchmark chart, published 2026-08-12 with the GA. A score is a +model-and-harness result, not a model-only one, and no independent same-harness +run of 0813 exists yet. Treat these as the claim, not the verdict. +
- No GPT-5.6 column. DeepSeek's chart compares against Kimi +K3, GLM-5.2, Claude Opus 4.8 and Fable 5 only. GPT figures elsewhere on this +page are drawn from those vendors' own releases and are cross-vendor, so they +are looser still. +
- Kimi and GLM are single-sourced. Those two columns appear +only on the extended variant of the chart; the shared columns are identical +across every copy, so the numbers are consistent, but the Kimi and GLM rows +rest on one source. +
- Independent history says be careful. The one held-out +check on the V4-Pro preview, from NIST/CAISI, put it closer to GPT-5 +and roughly eight months behind the frontier, below its self-reported +position. No equivalent 0813 evaluation exists yet. +
The launch table
+Higher is better. HLE is shown as without-tools / with-tools. Every figure +is from DeepSeek's GA chart of 2026-08-12; a dash means the vendor did not +report it.
+| Benchmark | +V4-Pro 0813 | +V4-Flash 0731 | +Kimi K3 | +GLM-5.2 | +Opus 4.8 | +Fable 5 | +
|---|---|---|---|---|---|---|
| Terminal-Bench 2.1 | 87.9 | 82.7 | 88.3 | 81.6 | 85.0 | 88.0 |
| DeepSWE | 62.7 | 54.4 | 67.5 | 46.2 | 58.0 | 70.0 |
| Toolathlon-Verified | 74.1 | 70.3 | 76.5 | 59.9 | 76.2 | 77.9 |
| CyberGym | 83.3 | 76.7 | 80.0 | – | 78.3 | 83.1 |
| NL2Repo | 61.5 | 54.2 | – | 48.9 | 69.7 | – |
| AutomationBench | 31.8 | 25.1 | 30.8 | 12.9 | 27.2 | 29.1 |
| DSBench-FullStack | 71.1 | 68.7 | 63.0 | 51.8 | 71.6 | 77.2 |
| DSBench-Hard | 67.2 | 59.6 | 73.7 | 54.5 | 71.7 | 68.3 |
| Agents' Last Exam | 25.7 | 25.2 | 24.5 | 23.9 | 25.7 | – |
| HLE (no / tools) | 42.7/60.0 | 37.8/51.5 | 43.5/56.0 | 40.5/54.7 | 49.8/57.9 | 53.3/63.0 |
The honest read of the row-by-row: V4-Pro is within a point of Fable +5 on Terminal-Bench and CyberGym, ahead of Opus 4.8 on +Terminal-Bench, DeepSWE, CyberGym and AutomationBench, and behind Kimi +K3 on Terminal-Bench, DeepSWE, Toolathlon and DSBench-Hard. It is in +the cluster, not clear of it. On knowledge without tools (HLE) the closed +models still lead. Flash trails Pro across the board but stays remarkably close +for a fifth of the price, which is the whole point of the next two sections.
+ +What GA changed
+The model ID stayed deepseek-v4-pro and the
+rate card did not move. What moved is the
+checkpoint. Against the April preview, DeepSeek's own chart shows the gains
+landing almost entirely in the agentic and SWE suites:
| Benchmark | Preview | GA 0813 | Delta |
|---|---|---|---|
| DeepSWE | 12.8 | 62.7 | +49.9 |
| DSBench-Hard | 31.1 | 67.2 | +36.1 |
| CyberGym | 52.7 | 83.3 | +30.6 |
| DSBench-FullStack | 41.8 | 71.1 | +29.3 |
| NL2Repo | 38.5 | 61.5 | +23.0 |
| Toolathlon-Verified | 55.9 | 74.1 | +18.2 |
| Terminal-Bench 2.1 | 72.1 | 87.9 | +15.8 |
A near five-fold jump on DeepSWE is not a new base model. Gains shaped like +this come from agent post-training: better tool-error recovery, better context +policy, reinforcement on the harness the benchmark runs in. That is real and it +is useful, and it is also exactly the kind of gain that can be harness-specific, +which is why the caveats above matter and why the preview numbers are the floor, +not these.
+ +The kill line (斩杀线)
+The term comes from Chinese gaming: the 斩杀线 is the +health threshold below which a target can be executed outright. Applied to +models, the idea is that DeepSeek's price-to-capability ratio draws a line, and +any model that is both weaker and more expensive falls below it and has +no reason to be chosen. The lever is price, and the gap is not small:
+| Per 1M tokens | v4-flash | v4-pro | GPT-5.6 Sol | pro is cheaper by |
|---|---|---|---|---|
| input, cache miss | $0.14 | $0.435 | $5.00 | ~11.5x |
| input, cache hit | $0.0028 | $0.003625 | $0.50 | ~138x |
| output | $0.28 | $0.87 | $30.00 | ~34.5x |
DeepSeek prices are the published USD rate card of +2026-08-02, unchanged at GA; they are a conversion of the RMB card +(¥3 / ¥0.025 / ¥6 per 1M for pro) at one consistent rate. GPT-5.6 +Sol prices are from OpenAI's own listing. A broad +DeepSeek repricing is announced but has no date, so it is not applied here.
+The sober version matters as much as the slogan. The kill line is real for +the middle of the market: a model that costs more than V4-Pro and +scores below it on the table above is hard to justify, and that is most of the +field. It is not real for the frontier. On the hardest +multi-step work the strongest closed models still finish in fewer turns and +need less steering, and per-attempt reliability is a thing you can measure in +wall-clock and interventions, not just in dollars. DeepSeek does not have to win +per attempt to win per dollar, and it does not have to win per dollar to lose +the one task where getting it right the first time is the whole job.
+ +What it means in practice
+The economics only pay off if the workflow is built for them. Three moves, +each of which this CLI is shaped to support:
+-
+
- Route by role. Use
deepseek-v4-profor +planning, ambiguous changes, security review and recovery; let +deepseek-v4-flashdo bounded implementations and parallel work. +ds chat -m deepseek-v4-proand the +Anthropic remap make the switch one flag.
+ - Structure prompts for the cache. A cached input token
+costs about 1/50th of an uncached one. Keep the system prompt, tool schemas,
+repository map and durable instructions in an identical prefix and put the
+volatile part last;
ds usagereports what the cache saved so you +can see whether it is working. This is where the 138x cache-hit number turns +from a table cell into a bill.
+ - Measure successful-task cost, not per-token price. A +cheaper model that retries five times can cost more than a dearer one that lands +first. The ledger stores exact token counts +per call, so you can price a whole task under any rate card, including the one +that has not been announced yet. +
Sources
+The launch table and the preview deltas are transcribed from DeepSeek's +official agent-benchmark chart, published on the +Models & Pricing page and +circulated on 2026-08-12; the extended variant carrying the Kimi K3 and GLM-5.2 +columns was the widest copy available. Rate-card figures are DeepSeek's own and +OpenAI's own. The independent-evaluation caveat refers to the NIST/CAISI review +of the V4-Pro preview. Numbers change fast and vendor charts are vendor charts; +cross-check a live leaderboard before betting on a single cell.
+