ERNIE 5.1
01
Snapshot
Overall
66
rank #17 of 22
Coding
62
rank #18 of 22
Multimodal
74
rank #16 of 22
Input / Output
$1.1 / $3.3
USD per 1M tokens
Context
128K
usable recall 82
Speed
~45 tok/s
TTFT 0.7s
Aggregated from public sources and independently weighted; methodology on the Terms page. Scores within 3 points are treated as statistically tied.
02
Specs & pricing
Every tracked field. List API prices in USD per 1M tokens; verify current vendor pricing before purchase.
| Vendor | Baidu (CN) |
| Released / current | 2026.05 |
| Licensing | Closed / managed API |
| Context window | 128K |
| Max output | 32K |
| Effective-context score | 82 |
| Input / output price | $1.1 / $3.3 per 1M tokens |
| Cache discount | No published cache discount |
| Free tier | ERNIE Bot free; quota on the Qianfan platform |
| Speed / TTFT | ~45 tok/s / 0.7s |
| Function calling | 78 |
| Refusal rate | ~12% |
| English / Chinese | 65 / 92 |
| Modalities | text, image |
| Fine-tuning | Yes |
| Private deployment | Yes (enterprise) |
| SOC2 / no-train | no / yes |
03
Strengths & watch-outs
Strengths
- Optimized for Chinese
- Search augmentation
- Baidu integration
- Mature enterprise services
Watch-outs
- Mid overall performance
- Weak English
- Small context
04
Where it ranks among 22 models
Overall 66/100 (#17), coding 62 (#18), multimodal 74 (#16). See the full boards and side-by-side compare on the leaderboard.
05
Pricing in practice
Illustrative monthly bill for 100M input + 30M output tokens: $209 at list. Your mix and cache-hit ratio change this.
Illustrative model using public list rates before any enterprise agreement; cache discount: No published cache discount.
06
Frequently asked questions
How much does ERNIE 5.1 cost?
List API pricing is $1.1 input and $3.3 output per 1M tokens. No published cache discount. ERNIE Bot free; quota on the Qianfan platform
What is ERNIE 5.1's context window, and is it usable end to end?
It advertises a 128K window with 32K max output; its effective-context score is 82/100, which is the better predictor of whether long-document details are actually retained.
Is ERNIE 5.1 good for coding?
Its coding aggregate is 62/100, rank #18 of 22. It is adequate for scripts and assisted completion rather than the most demanding SWE-style work.
Can I self-host ERNIE 5.1?
No. It is a closed managed API (private deployment may be negotiated for enterprise; there is no weights download.
How strong is it in Chinese and on multimodal inputs?
Chinese score is 92 versus English 65; multimodal aggregate is 74/100 (rank #16), accepting text, image.