sambanova API key fails test

[lesshaste]

I have a sambanova key which I can test successfully using curl. However the free-coding-models test fails. This is because it is using the first model in the list which is Qwen3-235B-A22B-Instruct-2507, but this model doesn't exist in the API.

To find out which models it has, I ran:

curl -sS 'https://api.sambanova.ai/v1/models' \
  -H "Authorization: Bearer $SAMBANOVA_API_KEY" |
jq -r '.data[].id' |
while IFS= read -r model; do
  code=$(curl -sS -o /dev/null -w '%{http_code}' \
    'https://api.sambanova.ai/v1/chat/completions' \
    -H 'Content-Type: application/json' \
    -H "Authorization: Bearer $SAMBANOVA_API_KEY" \
    --data "$(jq -nc --arg model "$model" '{model:$model,messages:[{role:"user",content:"Reply with OK"}]}')")

  [ "$code" = "200" ] && echo "$model: PASS" || echo "$model: FAIL (HTTP $code)"
done

This tests all the models https://api.sambanova.ai/v1/models claims to have. Many of them fail the test. The ones that currently pass are:

1. DeepSeek-V3-0324
2. DeepSeek-V3.1-Terminus
3. DeepSeek-V3.2
4. Llama-3.3-Swallow-70B-Instruct-v0.4
5. Meta-Llama-3.1-8B-Instruct
6. Qwen3-235B
7. gpt-oss-120b

[vava-nessa]

fixed it, can you try it ?

[lesshaste]

I think so. It says the test fails but I think that is due to rate limiting.

[vava-nessa]

In fact, it is a problem I'm working at right now, to test api key, it pings endpoints, but if endpoints are overloaded, it says the api key test fails.. So I added retry to other models but if they're all overloaded then... still the same, i'm working on a better solution for that