UpstageSolar Pro 3
Solar Pro 3 is an Upstage proprietary reasoning model, released 27 Jan 2026. Its best configuration ranks #576 of 785 on the BenchLeader Index at 43.2 ±8.9, in the lower half. The ± is the point: 446 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in instruction following (70) and lowest in coding (20). At $0.262 per million tokens blended it is cheaper than most ranked models. Output speed of 137 tokens per second puts it faster than most, with a first answer in 15.8 s. Last measured 6 Apr 2026.stale: no new result in six months
- Blended price
- $0.262/M
- $0.150 in · $0.600 out
- Output speed
- 137 tok/s
- measured by Artificial Analysis
- First answer
- 16 s
- first token 0.40 s
- Context
- 131k
- Full answer
- 19 s
- median, reasoning included
- Cost per run
- $0.079
- one full Intelligence Index run
- Released
- 27 Jan 2026
- Overall index43
- Reasoning41
- Coding20
- Agents & tools38
- Knowledge37
- Instruction following70
- Long context40
- Composite38
Versions
Upstage has shipped 2 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| Solar Pro 4 | 6 Aug 2026 | 53.3 | #291 |
| Solar Pro 3this page | 27 Jan 2026 | 43.2 | #576 |
How its index has moved
9 Sept 2026 to 21 Oct 2026The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| GPQA Diamond (AA)not in index | 72.4%#280 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 10.3%#319 | Artificial Analysis | |
| CritPt | 0.0%#315 | Artificial Analysis |
Coding
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| SciCode | 25.5%#209 | SciCode | |
| SciCode (AA)not in index | 25.5%#218 | Artificial Analysis | |
| Terminal-Bench 4.0 (AA)not in index | 0.0%#152 | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in index | 12.0%#190 | Artificial Analysis |
Agents & tools
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Terminal-Bench Hard | 7.6%#241 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 86.3%#88 | Artificial Analysis | |
| τ³-Banking (AA) | 8.7%#155 | Artificial Analysis | |
| AutomationBenchnot in index | 0.5%#207 | Artificial Analysis | |
| GDP.pdf (AA)not in index | 1.6%#194 | Artificial Analysis |
Knowledge
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-Omniscience | -53.4#419 | Artificial Analysis | |
| AA-Omniscience: accuracynot in index | 18.5%#336 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 11.8%#387 | Artificial Analysis |
Instruction following
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| IFBench | 71.2%#59 | Artificial Analysis |
Long context
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-LCR | 32.3%#384 | Artificial Analysis | |
| MLCRnot in index | 0.6%#167 | Artificial Analysis |
Composite
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA Intelligence Index v4.3.2 | 7.8#455 | Artificial Analysis |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Upstage | 124 tok/s | 0.88 s | $0.150 | $0.600 | 131k | – |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0002 | – | 18.0 s |
| Summarize a 30-page report | 12,000 / 600 | $0.0022 | – | 20.2 s |
| Code edit | 6,000 / 1,500 | $0.0018 | – | 26.7 s |
| Agentic coding session | 60,000 / 4,000 | $0.011 | – | 44.9 s |
| Structured extraction | 2,000 / 200 | $0.0004 | – | 17.3 s |
See also
Questions people ask
- What are Solar Pro 3's benchmark scores?
- As of 24 Oct 2026, Solar Pro 3 scores 43.2 ±8.9 on the BenchLeader Index, strongest in instruction following (70) and weakest in coding (20). That rests on 8 independent benchmark results, and every score on this page links to its publisher.
- How does Solar Pro 3 rank?
- Solar Pro 3 is #576 of 785 ranked configurations on 24 Oct 2026. 446 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- How much does Solar Pro 3 cost per 1M tokens?
- As of 24 Oct 2026, Solar Pro 3 costs $0.150 per 1M input tokens and $0.600 per 1M output tokens, $0.262 blended (three input tokens per output token).
- When was Solar Pro 3 released?
- Upstage released Solar Pro 3 on 27 Jan 2026.
Data as of 24 Oct 2026. Compare with another model.
Cite as: BenchLeader, “Solar Pro 3: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/solar-pro-3.