Baidu DeepSeek V4.1 Flash review: tool calling on par with official, frequent no-available-channel pulls down answer quality and speed
In the 2026-11-03/04 APB Full run, Baidu's BFCL accuracy of 86.40% is on par with the official channel's 86.13%. But 975 HLE and AA-LCR question-instances got no complete answer, and 50 of 90 ten-way speed requests returned no_available_channel. Q is 38.00 and P is 37.77.
On 2026-11-03/04, DeepSeek V4.1 Flash 1017 was tested on six providers through AIHubMix with the APB Full suite, using the same questions, parameters and request scripts for each. This APB Profile covers Baidu only: what each section measured and how it came out.
The numbers below are scores, not a ranking or a provider recommendation. The four scores, Q, F, P and R, come first; the overall is a summary index, not a certification, and it has no confidence interval. The scoring rules are defined in AIHubMix Provider Benchmark Standard v2.0.
Scores across six providers

| Provider | Q quality | F function | P speed | K cache | R stability | Overall |
|---|---|---|---|---|---|---|
| ByteDance | 42.05 | 74.98 | 84.37 | 93.83 | 100 | 66.37 |
| Baidu (this post) | 38.00 | 88.39 | 37.77 | 68.85 | 100 | 56.48 |
| Alibaba Cloud | 41.75 | 89.78 | 88.58 | 99.37 | 100 | 73.38 |
| DeepSeek official channel | 42.45 | 94.00 | 86.14 | 98.38 | 100 | 73.76 |
| Fireworks | 41.42 | 71.78 | 79.51 | 60.68 | 100 | 64.10 |
| Pareto | 44.58 | 93.42 | 57.69 | 27.58 | 100 | 68.72 |
Providers are listed in a fixed order, not a ranking. K is a sub-score of P, accounting for 20% of it, and is listed separately for reference. The overall score is computed with the composition rules of AIHubMix Provider Benchmark Standard, which also take in the comparison with the official channel and evidence completeness; it is not a direct weighting of the four scores in the table. It is a summary index with no confidence interval. The DeepSeek official channel and Alibaba Cloud differ by only 0.37 overall, which is not a distinguishable difference.
At a glance
- Answer quality: Q 38.00, with BFCL 86.40%, HLE 9.36% and AA-LCR 32.80%. The gap to the official channel is −4.46 percentage points, with a 95% interval of −5.78 to −3.16 that excludes 0. HLE had 864 request failures and AA-LCR 100, almost all in window 1, counted as wrong.
- Function support: F 88.39; 148 of 174 counted checks passed. Neither the cache-read field nor a reasoning-token count was returned; the over-context and orphan tool-result cases both returned 200.
- Response speed: P 37.77. 50 of 90 ten-way requests returned
no_available_channel; with more than half invalid, first body token and TPOT were both scored at penalty values. 78 of 90 single-concurrency requests were valid. - Cache: K 68.85. The 8k hot requests hit 5 of 9, with 3 request failures; 32k hit 8 times. Hits are read in 1,024-token blocks (7,168 for 8k, 31,744 for 32k).
- Service stability: R 100. All 300 dedicated requests succeeded on the first attempt; channel failures during the answering and speed runs are counted in Q and P.
Test setup
| Item | Detail |
|---|---|
| Model | DeepSeek V4.1 Flash 1017 |
| Requested model ID | baidu-deepseek-v4.1-flash-1017 |
| Returned model name | deepseek-v4.1-flash-1017 |
| Access route | Through AIHubMix with a provider-specific model ID, not a direct connection to the provider's account |
| Parameters | Fixed APB v2 request profile, temperature 1.0, max_tokens 16,384, 300 s timeout, no automatic retry |
| Test profile | The five providers other than Pareto shared one fixed APB v2 profile for Q/P/K/R; F also checked its parameter contract |
| Time | 2026-11-03 18:07 – 11-04 16:33; cache requests 11-04 22:10–22:28; all UTC+8. All six providers shared one Mac mini, one network and one account. |
| Samples | Q 4,250 question-instances; F 180 checks; P 180 requests; R 300; K 30 |
Requests used the AIHubMix quick start route with a provider-specific model ID; the other providers' IDs are listed on the models page. The model's own parameters are described in DeepSeek's API documentation.
Q — answer quality: 38.00
Answer quality draws on three suites: 500 BFCL questions answered 3 times each, 250 HLE questions answered 10 times each, and 25 AA-LCR questions answered 10 times each. BFCL is scored by automatic rules; HLE and AA-LCR are judged right or wrong by the judge model deepseek-flash. All six providers received the same questions, and the three suites are pooled into one accuracy. Ten answers were removed because the question could not be scored, leaving 4,240 scored question-instances.
| Suite | Questions × rounds | Correct | Accuracy | Service failures |
|---|---|---|---|---|
| BFCL tool calls, auto-scored | 500 × 3 | 1,296/1,500 | 86.40% | 0 |
| HLE hard QA, judge-scored | 250 × 10 | 233/2,490 | 9.36% | 875 |
| AA-LCR long-doc reasoning, judge-scored | 25 × 10 | 82/250 | 32.80% | 100 |
| Total · 95% interval 36.23%–39.58% | — | 1,611/4,240 | 38.00% | 975 |
Service failures (failed requests or truncated output) stay in the denominator as wrong answers: HLE had 864 request failures (863 in window 1, 1 in window 2) and 11 outputs cut off at the max_tokens ceiling; AA-LCR had 100 request failures, all in window 1.
| BFCL category | Passed | Pass rate |
|---|---|---|
| simple_python: one function, right call and arguments | 336/360 | 93.33% |
| multiple: pick the right one of several functions | 269/300 | 89.67% |
| parallel: one function called several times at once | 268/300 | 89.33% |
| parallel_multiple: several functions at once | 206/240 | 85.83% |
| irrelevance: no relevant function, so no call | 113/150 | 75.33% |
| ahm_multi_turn: three-stage tool chain | 104/150 | 69.33% |
Each BFCL question was answered 3 times; the table sums the 3 rounds.
Compared with the other five


| Provider | BFCL | HLE | AA-LCR | Combined | Vs official |
|---|---|---|---|---|---|
| ByteDance | 85.20% | 15.34% | 49.20% | 42.05% | −0.40 (−1.46 to +0.66), not distinguishable |
| Baidu (this post) | 86.40% | 9.36% | 32.80% | 38.00% | −4.46 (−5.78 to −3.16), below official |
| Alibaba Cloud | 87.00% | 13.29% | 53.60% | 41.75% | −0.71 (−1.72 to +0.26), not distinguishable |
| DeepSeek official channel | 86.13% | 15.06% | 53.20% | 42.45% | Baseline |
| Fireworks | 85.53% | 14.38% | 46.00% | 41.42% | −1.04 (−2.14 to +0.05), not distinguishable |
| Pareto | 85.27% | 19.32% | 52.00% | 44.58% | +2.12 (−0.02 to +4.18), not distinguishable |
Vs official: the accuracy gap in percentage points, with the 95% interval in brackets; an interval that crosses 0 is recorded as not distinguishable. The HLE and AA-LCR judge, deepseek-flash, belongs to the same DeepSeek family as the tested model and was not calibrated, so its judgments may be lenient.
F — function support: 88.39
The function suite looks only at whether interface behaviour matches the protocol, not at whether the model answers well. Sixty cases in six groups ran once in each of three windows: 180 checks, judged against the format each provider documents. Checks that were not applicable, invalid or unobserved leave the denominator. Baidu passed 148 of 174 counted checks and scored 88.39 on the group weights.
| Case group | Weight | Passed | Excluded | Group score |
|---|---|---|---|---|
| Streaming and message protocol | 20% | 27/27 | 3 | 100.00 |
| Tool interface | 20% | 30/30 | — | 100.00 |
| Output constraints and parameters | 20% | 29/30 | — | 96.67 |
| Context and state | 15% | 26/30 | — | 86.67 |
| Errors and recovery | 15% | 21/30 | — | 70.00 |
| Metering and cache protocol | 10% | 15/27 | 3 | 55.56 |
| Total · weighted score | — | 148/174 | 6 | 88.39 |
The weighted score applies the group weights; it is not simply passed checks divided by 174.
| Failed case | Windows | Reason |
|---|---|---|
| Over context · context and state | 3/3 | Returned 200 instead of a parameter error |
| Malformed messages · errors and recovery | 3/3 | 400 with an empty body, no structured error |
| Orphan tool result · errors and recovery | 3/3 | Returned 200 instead of a parameter error |
| Malformed stream_options · errors and recovery | 3/3 | 400 with an empty body, no structured error |
| Cache-read field · metering and cache protocol | 3/3 | No cache-read field returned, so reads could not be confirmed |
| Reasoning-token metering · metering and cache protocol | 3/3 | No reasoning-token count returned |
| Tool definitions counted in input · metering and cache protocol | 3/3 | Input tokens did not rise when tool definitions were added |
| Input count stable with cache · metering and cache protocol | 3/3 | No cache-read field returned, so reads could not be confirmed |
| Sampling diversity · output constraints and parameters | 1/3 | Outputs at temperature 0 and 1 differed too little for the pre-registered bar |
| Isolation between requests · context and state | 1/3 | Step 1 did not reply only STORED, so isolation could not be checked |
“Windows” is how many of the 3 windows failed the check.
| Status | Case | Count | Reason |
|---|---|---|---|
| Unobserved | Streaming stop sequence | 3 | The request for this step was not sent after the previous step, so it could not be judged |
| Not applicable | Explicit cache_control | 3 | Registered in advance as unsupported by this interface |
These checks leave the denominator; they count as neither a pass nor a failure.
The malformed-messages and malformed-stream-options error cases failed on several providers: they returned 400 with an empty body, so the client gets no structured error explanation. Reasoning-token metering behaved differently from provider to provider.
Compared with the other five

| Provider | Passed | Failed | Excluded | Weighted score |
|---|---|---|---|---|
| ByteDance | 124 | 50 | 6 | 74.98 |
| Baidu (this post) | 148 | 26 | 6 | 88.39 |
| Alibaba Cloud | 155 | 22 | 3 | 89.78 |
| DeepSeek official channel | 137 | 10 | 33 | 94.00 |
| Fireworks | 122 | 55 | 3 | 71.78 |
| Pareto | 29 | 4 | 147 | 93.42 |
P — response speed: 37.77
Each request carried about 10,000 input tokens and asked for about 1,100 output tokens of body text. In each of three windows, 30 requests ran at single concurrency and 30 at 10-way concurrency: 180 requests in all.
| Item | Detail |
|---|---|
| First body token (TTFA) | Median 2.26 s single (valid-only 2.05 s); at 10-way over half were invalid, so the median counts as infinite (valid-only 1.86 s) |
| Generation speed (TPOT) | 3.48 ms/token single; 50.00 ms/token at 10-way (penalty value, over half invalid) |
| Effective throughput | 1,329.2 token/s at 10-way |
| Valid requests | 78/90 single (window 1: 18/30), 40/90 at 10-way (window 1: 0/30, window 3: 10/30); invalid: 62 returned 400 no_available_channel |
| Sub-scores | TTFA 0.00, TPOT 0.00, throughput 100.00; TTFA and TPOT take the lower tier, weighted 35% / 35% / 30%: response score 30.00 |
| P score | Response score 30.00 × 80% + cache K 68.85 × 20% = 37.77 |
Invalid requests are scored as infinite, so the medians in the table can be higher than valid-only medians. TPOT is estimated from streamed chunks, not measured per token on the hardware. Throughput is valid output tokens at 10-way concurrency divided by batch wall-clock time.
Compared with the other five

| Provider | Valid requests | TTFA single | TTFA 10-way | TPOT (slower) | Throughput tok/s | Response |
|---|---|---|---|---|---|---|
| ByteDance | 180/180 | 2.12 s | 2.10 s | 3.43 ms | 1,250 | 82.00 |
| Baidu (this post) | 118/180 | 2.26 s | ∞ | 50.00 ms | 1,329 | 30.00 |
| Alibaba Cloud | 180/180 | 1.34 s | 1.35 s | 4.69 ms | 1,314 | 85.88 |
| DeepSeek official channel | 165/180 | 1.87 s | 1.72 s | 2.84 ms | 1,546 | 83.08 |
| Fireworks | 179/180 | 1.63 s | 1.64 s | 3.29 ms | 1,505 | 84.22 |
| Pareto | 110/180 | 5.85 s | 6.13 s | 4.63 ms | 519 | 65.22 |
All six shared one machine, network and account. The three windows were scheduled provider by provider, so collection periods do not fully overlap; small differences should not be over-read. ∞ means more than half the requests were invalid, so the median counts as infinite.
K — cache: 68.85
Each group sends five requests. The first carries a fixed prefix of about 8k or 32k tokens and seeds the cache. Requests 2–4 keep the prefix and change only the final question, at gaps of 5 s, 30 s and 120 s. The fifth uses a new prefix as a cold control. Three windows × two lengths make 30 requests. K is the cache tokens read by hot requests divided by the tokens that should be reusable under the rules, with the two lengths weighted equally; K is 20% of P.
| Window | 1 · seed | 2 · +5 s | 3 · +30 s | 4 · +120 s | 5 · cold |
|---|---|---|---|---|---|
| 8k prefix · hot hits 5/9 · read 35,840 of 72,252 reusable · 49.60 | — | — | — | — | — |
| Window 1 | Failed | Failed | Failed | Not returned | Not returned |
| Window 2 | Not returned | 7,168 | 7,168 | Failed | Failed |
| Window 3 | Not returned | 7,168 | 7,168 | 7,168 | Not returned |
| 32k prefix · hot hits 8/9 · read 253,952 of 288,261 reusable · 88.10 | — | — | — | — | — |
| Window 1 | Not returned | Not returned | 31,744 | 31,744 | Not returned |
| Window 2 | Not returned | 31,744 | 31,744 | 31,744 | Not returned |
| Window 3 | Not returned | 31,744 | 31,744 | 31,744 | Not returned |
| K (8k and 32k weighted equally) | — | — | — | — | 68.85 |
Values are cache-read tokens reported by the server; “Not returned” means the field was absent from the response.
Hits are read in 1,024-token blocks: 7,168 tokens for 8k and 31,744 for 32k. On a miss, usage carries no prompt_tokens_details; the table shows “Not returned”, counted as 0. The failed requests (8k window 1, requests 1, 2 and 3; 8k window 2, requests 4 and 5) returned no_available_channel.
Compared with the other five

| Provider | 8k hits | 8k | 32k hits | 32k | K |
|---|---|---|---|---|---|
| ByteDance | 9/9 | 98.84 | 8/9 | 88.81 | 93.83 |
| Baidu (this post) | 5/9 | 49.60 | 8/9 | 88.10 | 68.85 |
| Alibaba Cloud | 9/9 | 98.83 | 9/9 | 99.90 | 99.37 |
| DeepSeek official channel | 9/9 | 97.25 | 9/9 | 99.51 | 98.38 |
| Fireworks | 3/9 | 32.81 | 8/9 | 88.54 | 60.68 |
| Pareto | 2/9 | 21.89 | 3/9 | 33.27 | 27.58 |
R — service stability: 100
The 300 requests fall into four types: short text, long input, tool calls and multi-turn history. They ran at single concurrency across three windows. Only whether the first attempt returned a valid, complete response is checked; answers are not graded and nothing is retried. Failures in the answer-quality and speed sections are counted in Q and P, not again in R.
| Request type | First-attempt success |
|---|---|
| Short text | 120/120 |
| 10k long input | 90/90 |
| Tool calls | 60/60 |
| Multi-turn history | 30/30 |
| Total · 95% interval 98.78%–100% | 300/300 |
All 300 requests returned a valid, complete response on the first attempt. R = 100 describes these 300 dedicated requests only; it does not mean the whole chain had no failures. The lower bound of the 95% interval is 98.78%: 300 successes out of 300 can only say the failure rate is likely below 1.22%.
Compared with the other five
| Provider | First-attempt success | 95% interval |
|---|---|---|
| ByteDance | 300/300 | 98.78%–100% |
| Baidu (this post) | 300/300 | 98.78%–100% |
| Alibaba Cloud | 300/300 | 98.78%–100% |
| DeepSeek official channel | 300/300 | 98.78%–100% |
| Fireworks | 300/300 | 98.78%–100% |
| Pareto | 300/300 | 98.78%–100% |
R covers only these 300 dedicated requests; failures in the answer-quality and speed sections are recorded in Q and P and are not counted again in R.
FAQ
What does the overall score of 56.48 represent?
It is a summary index computed with the composition rules of AIHubMix Provider Benchmark Standard, which also take in the comparison with the official channel and evidence completeness. It is not a direct weighting of Q, F, P and R, has no confidence interval and is not a certification.
Why were more than half of the ten-way speed requests invalid?
50 of the 90 returned no_available_channel. With more than half invalid, the median first-token time counts as infinite, and TTFA and TPOT are scored at their penalty values.
Does R 100 mean there were no channel failures?
No. R covers 300 dedicated requests, all first-attempt successes. The 62 no_available_channel responses in the speed run and the failures in the answering run are counted under P and Q.
Why is K 68.85 when the 32k prefix hit 8 of 9?
The two lengths weigh equally. The 8k prefix hit 5 of 9 hot requests and scored 49.60, while 32k hit 8 of 9 and scored 88.10; the average is 68.85.
Is Baidu's answer quality distinguishable from the official channel?
Yes. The combined Q gap is −4.46 percentage points, with a 95% interval of −5.78 to −3.16 that excludes 0. Among the five providers compared with the official channel, it is the only interval that excludes 0.
Which F checks were excluded from the denominator?
Six: three streaming stop-sequence checks that were unobserved because the request for that step was not sent after the previous step, and three explicit cache_control checks registered in advance as unsupported by this interface.
Keep reading: the DeepSeek V4.1 Flash series
- Baidu's BFCL accuracy is close to the official channel's; ByteDance's run follows tool calls and metering on the same suite: ByteDance DeepSeek V4.1 Flash review: answers indistinguishable from official, input counts don't match.
- Baidu's single-concurrency first body token of 2.26 s compares with 1.34 s here: Alibaba Cloud DeepSeek V4.1 Flash review: 1.34s first body token, all 18 hot requests hit cache, answers indistinguishable from official.
- Baidu's F 88.39 and Q gap are measured against this baseline, which scored F 94.00: DeepSeek official channel DeepSeek V4.1 Flash review: comparison baseline, 15 invalid requests at single concurrency.
- Baidu's 8k cache hit 5 of 9; Fireworks hit 3 of 9 and also returned empty stop replies: Fireworks DeepSeek V4.1 Flash review: stop requests return empty body, 8k cache hot requests hit 3 of 9.
- Pareto also lost many speed requests, 70 of 180, but in a different way: Pareto DeepSeek V4.1 Flash review: only 5 hits in 18 hot requests, answers indistinguishable from official.
Sources
- DeepSeek V4.1 Flash 1017 model page (AIHubMix)
- AIHubMix Provider Benchmark Standard v2.0 (AIHubMix docs)
- Quick start (AIHubMix docs)
- Models (AIHubMix)
- DeepSeek API documentation (DeepSeek)